Mari Egan, MD

Illustrative case

A 67-year-old woman with diabetes, hypertension, and hyperlipidemia comes to your office for an evaluation before undergoing a total hip arthroplasty. She is not taking a beta-blocker. Should you prescribe one?

Current guidelines from the American College of Cardiology Foundation (ACCF) and the American Heart Association (AHA) recommend starting beta-blockers to prevent cardiac events in patients about to undergo intermediate- or high-risk surgery or vascular surgery who have a history of inducible ischemia, coronary artery disease (CAD), or at least one risk factor for CAD.² However, the majority of the evidence for these guidelines, which were published in 2009 and are in the process of being updated, came from the DECREASE (Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography) trials, which have been discredited due to serious methodological flaws, including falsified descriptions of how outcomes were determined and fictitious databases.³ A new meta-analysis by Bouri et al¹ that excluded the DECREASE trials found that although preoperative beta-blockers reduce the rate of certain nonfatal outcomes, they increase the risk of death and stroke.

STUDY SUMMARY: Multiple RCTs find preop beta-blockers do more harm, than good

Bouri et al¹ conducted a meta-analysis of published RCTs evaluating preoperative beta-blockers vs placebo for patients undergoing noncardiac surgery. Of the 11 studies that met eligibility criteria, 2 were the discredited DECREASE trials. Thus, Bouri et al¹ analyzed 9 high-quality RCTs that included 10,529 patients.

Most studies included patients undergoing vascular surgery. Some studies also included intra-abdominal, intrathoracic, neurosurgical, orthopedic, urologic, and gynecologic surgeries. Beta-blockers were started no more than a day before surgery and were discontinued at hospital discharge or up to 30 days postop. Metoprolol was used in 5 trials, bisoprolol in one trial, atenolol in 2 trials, and propranolol in one trial. The primary endpoint was all-cause mortality within 30 days.

A total of 5264 patients were randomized to beta-blockers and 5265 to placebo. There were 162 deaths in the beta-blocker group and 129 deaths in the placebo group. Patients who received beta-blockers had a 27% increased risk of all-cause mortality (risk ratio [RR]=1.27; 95% confidence interval [CI], 1.01-1.60; P=.04). The number needed to harm was 160.

Six of the studies also evaluated rates of nonfatal myocardial infarction (MI), nonfatal stroke, and hypotension. Beta-blockers lowered the risk of nonfatal MI (RR=.73; 95% CI, .61-.88; P=.001), but increased the risk of nonfatal stroke (RR=1.73; 95% CI, 1.00-2.99; P=.05) and hypotension (RR=1.51; 95% CI, 1.37-1.67; P=.00001).

While beta-blockers protect against nonfatal MIs, they increase the risk of nonfatal strokes and death.This meta-analysis was dominated by the 2008 Peri-Operative ISchemic Evaluation (POISE) trial, an RCT that compared placebo to extended-release metoprolol, 100 mg 2 to 4 hours before surgery followed by 200 mg/d for 30 days, in 8351 patients with, or at risk for, atherosclerotic disease.⁴ While beta-blockers reduced the risk of MI and atrial fibrillation, they increased the risk of mortality and stroke, likely due to drug-induced hypotension. The slightly larger-than-typical doses of beta-blockers used in the POISE study may have contributed to the excess mortality.

WHAT'S NEW: Avoiding beta-blockers in surgery patients will prevent deaths

Bouri et al¹ found that while beta-blockers protect against nonfatal MIs, they increase the risk for nonfatal strokes and death. This new meta-analysis challenges the ACCF/AHA recommendations by suggesting that abandoning the use of beta-blockers for preoperative patients who aren’t already taking them will prevent a substantial number of perioperative deaths. Bouri et al¹ estimate that in the United Kingdom, where 47,286 deaths occur annually within 30 days of intermediate or high-risk procedures, the number of iatrogenic deaths would drop by approximately 10,000 if beta-blockers were not used.¹

CAVEATS: Don't stop beta-blockers in patients who already take them

This meta-analysis did not evaluate outcomes in patients who were already taking beta-blockers. Patients who are already on beta-blockers should continue to take them in the perioperative period, which is in line with current ACCF/AHA guidelines.

CHALLENGES TO IMPLEMENTATION: Some physician may be reluctant to disregard published guidelines 

Some physicians may not be comfortable ignoring the current ACCF/AHA guidelines that make a Class IIa recommendation (it is reasonable to administer this treatment) for the use of preoperative beta-blockade for patients at risk of cardiovascular events who were not previously taking a beta-blocker. This updated meta-analysis excludes the discredited DECREASE trials and challenges us to act against these current guidelines while we wait for updated recommendations.

Acknowledgement
The PURLs Surveillance System was supported in part by Grant Number UL1RR024999 from the National Center For Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center For Research Resources or the National Institutes of Health.

Illustrative case

A 67-year-old woman with diabetes, hypertension, and hyperlipidemia comes to your office for an evaluation before undergoing a total hip arthroplasty. She is not taking a beta-blocker. Should you prescribe one?

Current guidelines from the American College of Cardiology Foundation (ACCF) and the American Heart Association (AHA) recommend starting beta-blockers to prevent cardiac events in patients about to undergo intermediate- or high-risk surgery or vascular surgery who have a history of inducible ischemia, coronary artery disease (CAD), or at least one risk factor for CAD.² However, the majority of the evidence for these guidelines, which were published in 2009 and are in the process of being updated, came from the DECREASE (Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography) trials, which have been discredited due to serious methodological flaws, including falsified descriptions of how outcomes were determined and fictitious databases.³ A new meta-analysis by Bouri et al¹ that excluded the DECREASE trials found that although preoperative beta-blockers reduce the rate of certain nonfatal outcomes, they increase the risk of death and stroke.

STUDY SUMMARY: Multiple RCTs find preop beta-blockers do more harm, than good

Bouri et al¹ conducted a meta-analysis of published RCTs evaluating preoperative beta-blockers vs placebo for patients undergoing noncardiac surgery. Of the 11 studies that met eligibility criteria, 2 were the discredited DECREASE trials. Thus, Bouri et al¹ analyzed 9 high-quality RCTs that included 10,529 patients.

Most studies included patients undergoing vascular surgery. Some studies also included intra-abdominal, intrathoracic, neurosurgical, orthopedic, urologic, and gynecologic surgeries. Beta-blockers were started no more than a day before surgery and were discontinued at hospital discharge or up to 30 days postop. Metoprolol was used in 5 trials, bisoprolol in one trial, atenolol in 2 trials, and propranolol in one trial. The primary endpoint was all-cause mortality within 30 days.

A total of 5264 patients were randomized to beta-blockers and 5265 to placebo. There were 162 deaths in the beta-blocker group and 129 deaths in the placebo group. Patients who received beta-blockers had a 27% increased risk of all-cause mortality (risk ratio [RR]=1.27; 95% confidence interval [CI], 1.01-1.60; P=.04). The number needed to harm was 160.

Six of the studies also evaluated rates of nonfatal myocardial infarction (MI), nonfatal stroke, and hypotension. Beta-blockers lowered the risk of nonfatal MI (RR=.73; 95% CI, .61-.88; P=.001), but increased the risk of nonfatal stroke (RR=1.73; 95% CI, 1.00-2.99; P=.05) and hypotension (RR=1.51; 95% CI, 1.37-1.67; P=.00001).

While beta-blockers protect against nonfatal MIs, they increase the risk of nonfatal strokes and death.This meta-analysis was dominated by the 2008 Peri-Operative ISchemic Evaluation (POISE) trial, an RCT that compared placebo to extended-release metoprolol, 100 mg 2 to 4 hours before surgery followed by 200 mg/d for 30 days, in 8351 patients with, or at risk for, atherosclerotic disease.⁴ While beta-blockers reduced the risk of MI and atrial fibrillation, they increased the risk of mortality and stroke, likely due to drug-induced hypotension. The slightly larger-than-typical doses of beta-blockers used in the POISE study may have contributed to the excess mortality.

WHAT'S NEW: Avoiding beta-blockers in surgery patients will prevent deaths

Bouri et al¹ found that while beta-blockers protect against nonfatal MIs, they increase the risk for nonfatal strokes and death. This new meta-analysis challenges the ACCF/AHA recommendations by suggesting that abandoning the use of beta-blockers for preoperative patients who aren’t already taking them will prevent a substantial number of perioperative deaths. Bouri et al¹ estimate that in the United Kingdom, where 47,286 deaths occur annually within 30 days of intermediate or high-risk procedures, the number of iatrogenic deaths would drop by approximately 10,000 if beta-blockers were not used.¹

CAVEATS: Don't stop beta-blockers in patients who already take them

This meta-analysis did not evaluate outcomes in patients who were already taking beta-blockers. Patients who are already on beta-blockers should continue to take them in the perioperative period, which is in line with current ACCF/AHA guidelines.

CHALLENGES TO IMPLEMENTATION: Some physician may be reluctant to disregard published guidelines 

Some physicians may not be comfortable ignoring the current ACCF/AHA guidelines that make a Class IIa recommendation (it is reasonable to administer this treatment) for the use of preoperative beta-blockade for patients at risk of cardiovascular events who were not previously taking a beta-blocker. This updated meta-analysis excludes the discredited DECREASE trials and challenges us to act against these current guidelines while we wait for updated recommendations.

Acknowledgement
The PURLs Surveillance System was supported in part by Grant Number UL1RR024999 from the National Center For Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center For Research Resources or the National Institutes of Health.

Illustrative case

A 67-year-old woman with diabetes, hypertension, and hyperlipidemia comes to your office for an evaluation before undergoing a total hip arthroplasty. She is not taking a beta-blocker. Should you prescribe one?

Current guidelines from the American College of Cardiology Foundation (ACCF) and the American Heart Association (AHA) recommend starting beta-blockers to prevent cardiac events in patients about to undergo intermediate- or high-risk surgery or vascular surgery who have a history of inducible ischemia, coronary artery disease (CAD), or at least one risk factor for CAD.² However, the majority of the evidence for these guidelines, which were published in 2009 and are in the process of being updated, came from the DECREASE (Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography) trials, which have been discredited due to serious methodological flaws, including falsified descriptions of how outcomes were determined and fictitious databases.³ A new meta-analysis by Bouri et al¹ that excluded the DECREASE trials found that although preoperative beta-blockers reduce the rate of certain nonfatal outcomes, they increase the risk of death and stroke.

STUDY SUMMARY: Multiple RCTs find preop beta-blockers do more harm, than good

Bouri et al¹ conducted a meta-analysis of published RCTs evaluating preoperative beta-blockers vs placebo for patients undergoing noncardiac surgery. Of the 11 studies that met eligibility criteria, 2 were the discredited DECREASE trials. Thus, Bouri et al¹ analyzed 9 high-quality RCTs that included 10,529 patients.

Most studies included patients undergoing vascular surgery. Some studies also included intra-abdominal, intrathoracic, neurosurgical, orthopedic, urologic, and gynecologic surgeries. Beta-blockers were started no more than a day before surgery and were discontinued at hospital discharge or up to 30 days postop. Metoprolol was used in 5 trials, bisoprolol in one trial, atenolol in 2 trials, and propranolol in one trial. The primary endpoint was all-cause mortality within 30 days.

A total of 5264 patients were randomized to beta-blockers and 5265 to placebo. There were 162 deaths in the beta-blocker group and 129 deaths in the placebo group. Patients who received beta-blockers had a 27% increased risk of all-cause mortality (risk ratio [RR]=1.27; 95% confidence interval [CI], 1.01-1.60; P=.04). The number needed to harm was 160.

Six of the studies also evaluated rates of nonfatal myocardial infarction (MI), nonfatal stroke, and hypotension. Beta-blockers lowered the risk of nonfatal MI (RR=.73; 95% CI, .61-.88; P=.001), but increased the risk of nonfatal stroke (RR=1.73; 95% CI, 1.00-2.99; P=.05) and hypotension (RR=1.51; 95% CI, 1.37-1.67; P=.00001).

While beta-blockers protect against nonfatal MIs, they increase the risk of nonfatal strokes and death.This meta-analysis was dominated by the 2008 Peri-Operative ISchemic Evaluation (POISE) trial, an RCT that compared placebo to extended-release metoprolol, 100 mg 2 to 4 hours before surgery followed by 200 mg/d for 30 days, in 8351 patients with, or at risk for, atherosclerotic disease.⁴ While beta-blockers reduced the risk of MI and atrial fibrillation, they increased the risk of mortality and stroke, likely due to drug-induced hypotension. The slightly larger-than-typical doses of beta-blockers used in the POISE study may have contributed to the excess mortality.

WHAT'S NEW: Avoiding beta-blockers in surgery patients will prevent deaths

Bouri et al¹ found that while beta-blockers protect against nonfatal MIs, they increase the risk for nonfatal strokes and death. This new meta-analysis challenges the ACCF/AHA recommendations by suggesting that abandoning the use of beta-blockers for preoperative patients who aren’t already taking them will prevent a substantial number of perioperative deaths. Bouri et al¹ estimate that in the United Kingdom, where 47,286 deaths occur annually within 30 days of intermediate or high-risk procedures, the number of iatrogenic deaths would drop by approximately 10,000 if beta-blockers were not used.¹

CAVEATS: Don't stop beta-blockers in patients who already take them

This meta-analysis did not evaluate outcomes in patients who were already taking beta-blockers. Patients who are already on beta-blockers should continue to take them in the perioperative period, which is in line with current ACCF/AHA guidelines.

CHALLENGES TO IMPLEMENTATION: Some physician may be reluctant to disregard published guidelines 

Some physicians may not be comfortable ignoring the current ACCF/AHA guidelines that make a Class IIa recommendation (it is reasonable to administer this treatment) for the use of preoperative beta-blockade for patients at risk of cardiovascular events who were not previously taking a beta-blocker. This updated meta-analysis excludes the discredited DECREASE trials and challenges us to act against these current guidelines while we wait for updated recommendations.

Acknowledgement
The PURLs Surveillance System was supported in part by Grant Number UL1RR024999 from the National Center For Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center For Research Resources or the National Institutes of Health.

ILLUSTRATIVE CASE

A 43-year-old woman with a body mass index (BMI) of 38 kg/m² and a 5-year history of diabetes has a glycated hemoglobin (HbA1c) of 8.5% despite the use of oral hypoglycemic agents. Should you talk to her about gastric bypass surgery to treat her diabetes?

Diet and exercise are the first steps in addressing diabetes, but these interventions are often unsuccessful. The International Diabetes Federation (IDF) recommends consideration of bariatric surgery for patients who have a BMI >35 kg/m² and diabetes that lifestyle modification and pharmacotherapy have failed to control.²

Surgery for diabetes: Is there ample evidence?
Until recently, the IDF’s recommendation was based on observational data and a single RCT that found increased resolution of diabetes following various bariatric procedures.^3-5 In the study detailed in this PURL, Mingrone et al took another look.

STUDY SUMMARY: Surgery led to higher remission rates

This single-center, nonblinded RCT compared 2 malabsorptive procedures— Roux-en-Y gastric bypass and biliopancreatic diversion, a more complicated procedure not commonly performed—with medical therapy.¹ The primary outcome was the rate of diabetes remission at 2 years, defined as a fasting glucose level <100 mg/dL and an HbA1c <6.5%. Changes in BMI and cholesterol levels were among the secondary endpoints.

To be eligible, patients had to be between the ages of 30 and 60 years and have a BMI ≥35 kg/m², a history of type 2 diabetes ≥5 years, and an HbA1c ≥7.0%. Exclusion criteria included a history of type 1 diabetes, diabetes caused by an underlying disease or steroid treatment, previous bariatric surgery, pregnancy, diabetic complications, other severe medical conditions, and acute hospitalization. Both the gastric bypass and biliopancreatic diversion procedures were performed by independent surgical teams.

Participants (N=60) were evaluated at baseline and at 1, 3, 6, 9, 12, and 24 months after the intervention by a team that included a dietician, a nurse, and a physician. All received a diet plan with daily exercise designed by their team. Those in the medical therapy group had their medications titrated to reach a goal HbA1c <7%. Pharmacotherapy was stopped based on normalization of blood sugars or HbA1c <6.5%.

Within 15 days postsurgery, patients in both surgical arms had their diabetes medications stopped based on their blood glucose levels.

At 2 years, 75% of the patients in the gastric bypass arm and 95% of the patients in the biliopancreatic diversion arm (number needed to treat=1.3 and 1, respectively) were considered to be in diabetes remission, defined as a fasting blood sugar of <100 mg/dL and an HbA1c <6.5% after one year without pharmacotherapy. (Notably, this differs from that of the American Diabetes Association, which requires an HbA1c <6.0% for classification as complete remission.) None of the patients in the medical therapy arm was in remission at the 2-year mark.

On average, blood sugars normalized for gastric bypass patients by 10±2 months, vs 4±1 months for biliopancreatic diversion patients (P=.01). The average HbA1c at the end of 2 years was significantly different among all 3 groups (6.35%±1.42 for those undergoing gastric bypass, 4.95%±0.49 for the biliopancreatic diversion group, and 7.69%±0.57 for the medical therapy group), as was the change in HbA1c from baseline (TABLE). Changes in BMI and the number of patients who achieved normalization of total cholesterol were similar for both surgical groups. Interestingly, neither baseline BMI nor amount of weight lost or pre-enrollment duration of diabetes were predictors of diabetes remission or normalization of fasting glucose levels.

TABLE
Surgery vs medical therapy for diabetes: Gastric bypass and biliopancreatic diversion are more effective

There were no deaths associated with this study. There were 2 adverse events requiring reoperation: an incisional hernia in a patient in the biliopancreatic diversion group and an intestinal obstruction in a patient in the gastric bypass group. Six patients in the biliopancreatic diversion arm developed metabolic abnormalities, including iron deficiency anemia, hypoalbuminemia, osteopenia, and osteoporosis. In the gastric bypass arm, 2 patients developed iron deficiency anemia.

WHAT’S NEW?: Evidence of efficacy has grown

This is the first RCT to evaluate biliopancreatic diversion and only the second to evaluate gastric bypass as strategies for controlling diabetes. Similar findings were demonstrated at 12 months in an RCT of 150 obese patients with diabetes in which intensive medical therapy was compared with either gastric bypass or sleeve gastrectomy,⁶ published concurrently with the Mingrone study. Like the Mingrone study, this study found that for select patients with diabetes, surgery may lead to better outcomes than medical management alone.

CAVEATS: Long-term effect is still uncertain

The long-term efficacy of surgery as a way to manage diabetes remains uncertain. Patients in this study were followed for just 2 years and the outcomes were metabolic measures rather than morbidity and mortality. A recent prospective observational study following patients for 6 years after gastric bypass found that the rate of remission for diabetes was 75% (95% confidence interval (CI), 63%-87%) at 2 years but dropped to 62% (95% CI, 49%-75%) at 6 years⁷

A larger study (N=4047) of longer duration—the Swedish Obese Subjects (SOS) cohort study —found a considerably larger drop: The diabetes remission rate for those who had surgery went from 72% at 2 years to 36% at 10-year follow-up, but that was still higher than the 10-year remission rate (13%) for the matched controls.⁴ It is still not clear exactly how long diabetic remission lasts after bariatric surgery or what effect a 10-year respite from the disease will have on the long-term morbidity and mortality of patients with diabetes.

Surgical risks. In small studies such as the one by Mingrone et al,¹ it can be difficult to see the full extent of surgical complications. The much larger SOS study found low mortality rates (0.25%). But 13% of those who underwent bariatric surgery had postoperative complications (number needed to harm = 8), with 2.2% of patients requiring reoperation.⁴ Additionally, women who become pregnant after bariatric surgery are at increased risk for internal hernias or bowel obstruction during pregnancy.⁸

Furthermore, malabsorptive-type surgeries are known to cause nutritional deficiencies, leading to disorders including anemia and osteoporosis.⁶ Importantly, while women of childbearing-age who undergo bariatric surgery decrease their risk of developing gestational hypertension and gestational diabetes, they are more likely to have nutritional deficiencies during pregnancy and to have children with these deficiencies.⁸

CHALLENGES TO IMPLEMENTATION: The ideal candidate remains unclear

It is still not clear from this study which patients should be referred for bariatric surgery. Historically, BMI has been used as the main indication for bariatric surgery, but this and other, studies have found that remission of diabetes is independent of BMI and the amount of weight lost.⁹ A predictive 10-point Diabetes Surgery Score has recently been developed: It uses age, BMI, duration of diabetes, and C-peptide levels to predict the likelihood of diabetes remission after surgery.¹⁰ This scoring system has yet to be validated in non-Asian patients, and a threshold for recommending surgery has been not established. However, this tool indicates that younger patients with a shorter duration of diabetes (which was not a factor in the outcome of the Mingrone study) and no baseline use of insulin are most likely to benefit from surgery. Thus, these patients may be the ones we need to consider referring first.

Cost of surgery. Several studies have shown that bariatric surgery is cost-effective for the treatment of diabetes, and saves money after approximately 5 years.^11,12 However, patients with diabetes and obesity may be uninsured or underinsured, and have high out-of-pocket costs. One challenge will be to ensure that surgery is a viable option for patients with financial constraints.

Acknowledgement

The PURLs Surveillance System was supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

ILLUSTRATIVE CASE

A 43-year-old woman with a body mass index (BMI) of 38 kg/m² and a 5-year history of diabetes has a glycated hemoglobin (HbA1c) of 8.5% despite the use of oral hypoglycemic agents. Should you talk to her about gastric bypass surgery to treat her diabetes?

Diet and exercise are the first steps in addressing diabetes, but these interventions are often unsuccessful. The International Diabetes Federation (IDF) recommends consideration of bariatric surgery for patients who have a BMI >35 kg/m² and diabetes that lifestyle modification and pharmacotherapy have failed to control.²

Surgery for diabetes: Is there ample evidence?
Until recently, the IDF’s recommendation was based on observational data and a single RCT that found increased resolution of diabetes following various bariatric procedures.^3-5 In the study detailed in this PURL, Mingrone et al took another look.

STUDY SUMMARY: Surgery led to higher remission rates

This single-center, nonblinded RCT compared 2 malabsorptive procedures— Roux-en-Y gastric bypass and biliopancreatic diversion, a more complicated procedure not commonly performed—with medical therapy.¹ The primary outcome was the rate of diabetes remission at 2 years, defined as a fasting glucose level <100 mg/dL and an HbA1c <6.5%. Changes in BMI and cholesterol levels were among the secondary endpoints.

To be eligible, patients had to be between the ages of 30 and 60 years and have a BMI ≥35 kg/m², a history of type 2 diabetes ≥5 years, and an HbA1c ≥7.0%. Exclusion criteria included a history of type 1 diabetes, diabetes caused by an underlying disease or steroid treatment, previous bariatric surgery, pregnancy, diabetic complications, other severe medical conditions, and acute hospitalization. Both the gastric bypass and biliopancreatic diversion procedures were performed by independent surgical teams.

Participants (N=60) were evaluated at baseline and at 1, 3, 6, 9, 12, and 24 months after the intervention by a team that included a dietician, a nurse, and a physician. All received a diet plan with daily exercise designed by their team. Those in the medical therapy group had their medications titrated to reach a goal HbA1c <7%. Pharmacotherapy was stopped based on normalization of blood sugars or HbA1c <6.5%.

Within 15 days postsurgery, patients in both surgical arms had their diabetes medications stopped based on their blood glucose levels.

At 2 years, 75% of the patients in the gastric bypass arm and 95% of the patients in the biliopancreatic diversion arm (number needed to treat=1.3 and 1, respectively) were considered to be in diabetes remission, defined as a fasting blood sugar of <100 mg/dL and an HbA1c <6.5% after one year without pharmacotherapy. (Notably, this differs from that of the American Diabetes Association, which requires an HbA1c <6.0% for classification as complete remission.) None of the patients in the medical therapy arm was in remission at the 2-year mark.

On average, blood sugars normalized for gastric bypass patients by 10±2 months, vs 4±1 months for biliopancreatic diversion patients (P=.01). The average HbA1c at the end of 2 years was significantly different among all 3 groups (6.35%±1.42 for those undergoing gastric bypass, 4.95%±0.49 for the biliopancreatic diversion group, and 7.69%±0.57 for the medical therapy group), as was the change in HbA1c from baseline (TABLE). Changes in BMI and the number of patients who achieved normalization of total cholesterol were similar for both surgical groups. Interestingly, neither baseline BMI nor amount of weight lost or pre-enrollment duration of diabetes were predictors of diabetes remission or normalization of fasting glucose levels.

TABLE
Surgery vs medical therapy for diabetes: Gastric bypass and biliopancreatic diversion are more effective

There were no deaths associated with this study. There were 2 adverse events requiring reoperation: an incisional hernia in a patient in the biliopancreatic diversion group and an intestinal obstruction in a patient in the gastric bypass group. Six patients in the biliopancreatic diversion arm developed metabolic abnormalities, including iron deficiency anemia, hypoalbuminemia, osteopenia, and osteoporosis. In the gastric bypass arm, 2 patients developed iron deficiency anemia.

WHAT’S NEW?: Evidence of efficacy has grown

This is the first RCT to evaluate biliopancreatic diversion and only the second to evaluate gastric bypass as strategies for controlling diabetes. Similar findings were demonstrated at 12 months in an RCT of 150 obese patients with diabetes in which intensive medical therapy was compared with either gastric bypass or sleeve gastrectomy,⁶ published concurrently with the Mingrone study. Like the Mingrone study, this study found that for select patients with diabetes, surgery may lead to better outcomes than medical management alone.

CAVEATS: Long-term effect is still uncertain

The long-term efficacy of surgery as a way to manage diabetes remains uncertain. Patients in this study were followed for just 2 years and the outcomes were metabolic measures rather than morbidity and mortality. A recent prospective observational study following patients for 6 years after gastric bypass found that the rate of remission for diabetes was 75% (95% confidence interval (CI), 63%-87%) at 2 years but dropped to 62% (95% CI, 49%-75%) at 6 years⁷

A larger study (N=4047) of longer duration—the Swedish Obese Subjects (SOS) cohort study —found a considerably larger drop: The diabetes remission rate for those who had surgery went from 72% at 2 years to 36% at 10-year follow-up, but that was still higher than the 10-year remission rate (13%) for the matched controls.⁴ It is still not clear exactly how long diabetic remission lasts after bariatric surgery or what effect a 10-year respite from the disease will have on the long-term morbidity and mortality of patients with diabetes.

Surgical risks. In small studies such as the one by Mingrone et al,¹ it can be difficult to see the full extent of surgical complications. The much larger SOS study found low mortality rates (0.25%). But 13% of those who underwent bariatric surgery had postoperative complications (number needed to harm = 8), with 2.2% of patients requiring reoperation.⁴ Additionally, women who become pregnant after bariatric surgery are at increased risk for internal hernias or bowel obstruction during pregnancy.⁸

Furthermore, malabsorptive-type surgeries are known to cause nutritional deficiencies, leading to disorders including anemia and osteoporosis.⁶ Importantly, while women of childbearing-age who undergo bariatric surgery decrease their risk of developing gestational hypertension and gestational diabetes, they are more likely to have nutritional deficiencies during pregnancy and to have children with these deficiencies.⁸

CHALLENGES TO IMPLEMENTATION: The ideal candidate remains unclear

It is still not clear from this study which patients should be referred for bariatric surgery. Historically, BMI has been used as the main indication for bariatric surgery, but this and other, studies have found that remission of diabetes is independent of BMI and the amount of weight lost.⁹ A predictive 10-point Diabetes Surgery Score has recently been developed: It uses age, BMI, duration of diabetes, and C-peptide levels to predict the likelihood of diabetes remission after surgery.¹⁰ This scoring system has yet to be validated in non-Asian patients, and a threshold for recommending surgery has been not established. However, this tool indicates that younger patients with a shorter duration of diabetes (which was not a factor in the outcome of the Mingrone study) and no baseline use of insulin are most likely to benefit from surgery. Thus, these patients may be the ones we need to consider referring first.

Cost of surgery. Several studies have shown that bariatric surgery is cost-effective for the treatment of diabetes, and saves money after approximately 5 years.^11,12 However, patients with diabetes and obesity may be uninsured or underinsured, and have high out-of-pocket costs. One challenge will be to ensure that surgery is a viable option for patients with financial constraints.

Acknowledgement

The PURLs Surveillance System was supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

ILLUSTRATIVE CASE

A 43-year-old woman with a body mass index (BMI) of 38 kg/m² and a 5-year history of diabetes has a glycated hemoglobin (HbA1c) of 8.5% despite the use of oral hypoglycemic agents. Should you talk to her about gastric bypass surgery to treat her diabetes?

Diet and exercise are the first steps in addressing diabetes, but these interventions are often unsuccessful. The International Diabetes Federation (IDF) recommends consideration of bariatric surgery for patients who have a BMI >35 kg/m² and diabetes that lifestyle modification and pharmacotherapy have failed to control.²

Surgery for diabetes: Is there ample evidence?
Until recently, the IDF’s recommendation was based on observational data and a single RCT that found increased resolution of diabetes following various bariatric procedures.^3-5 In the study detailed in this PURL, Mingrone et al took another look.

STUDY SUMMARY: Surgery led to higher remission rates

This single-center, nonblinded RCT compared 2 malabsorptive procedures— Roux-en-Y gastric bypass and biliopancreatic diversion, a more complicated procedure not commonly performed—with medical therapy.¹ The primary outcome was the rate of diabetes remission at 2 years, defined as a fasting glucose level <100 mg/dL and an HbA1c <6.5%. Changes in BMI and cholesterol levels were among the secondary endpoints.

To be eligible, patients had to be between the ages of 30 and 60 years and have a BMI ≥35 kg/m², a history of type 2 diabetes ≥5 years, and an HbA1c ≥7.0%. Exclusion criteria included a history of type 1 diabetes, diabetes caused by an underlying disease or steroid treatment, previous bariatric surgery, pregnancy, diabetic complications, other severe medical conditions, and acute hospitalization. Both the gastric bypass and biliopancreatic diversion procedures were performed by independent surgical teams.

Participants (N=60) were evaluated at baseline and at 1, 3, 6, 9, 12, and 24 months after the intervention by a team that included a dietician, a nurse, and a physician. All received a diet plan with daily exercise designed by their team. Those in the medical therapy group had their medications titrated to reach a goal HbA1c <7%. Pharmacotherapy was stopped based on normalization of blood sugars or HbA1c <6.5%.

Within 15 days postsurgery, patients in both surgical arms had their diabetes medications stopped based on their blood glucose levels.

At 2 years, 75% of the patients in the gastric bypass arm and 95% of the patients in the biliopancreatic diversion arm (number needed to treat=1.3 and 1, respectively) were considered to be in diabetes remission, defined as a fasting blood sugar of <100 mg/dL and an HbA1c <6.5% after one year without pharmacotherapy. (Notably, this differs from that of the American Diabetes Association, which requires an HbA1c <6.0% for classification as complete remission.) None of the patients in the medical therapy arm was in remission at the 2-year mark.

On average, blood sugars normalized for gastric bypass patients by 10±2 months, vs 4±1 months for biliopancreatic diversion patients (P=.01). The average HbA1c at the end of 2 years was significantly different among all 3 groups (6.35%±1.42 for those undergoing gastric bypass, 4.95%±0.49 for the biliopancreatic diversion group, and 7.69%±0.57 for the medical therapy group), as was the change in HbA1c from baseline (TABLE). Changes in BMI and the number of patients who achieved normalization of total cholesterol were similar for both surgical groups. Interestingly, neither baseline BMI nor amount of weight lost or pre-enrollment duration of diabetes were predictors of diabetes remission or normalization of fasting glucose levels.

TABLE
Surgery vs medical therapy for diabetes: Gastric bypass and biliopancreatic diversion are more effective

There were no deaths associated with this study. There were 2 adverse events requiring reoperation: an incisional hernia in a patient in the biliopancreatic diversion group and an intestinal obstruction in a patient in the gastric bypass group. Six patients in the biliopancreatic diversion arm developed metabolic abnormalities, including iron deficiency anemia, hypoalbuminemia, osteopenia, and osteoporosis. In the gastric bypass arm, 2 patients developed iron deficiency anemia.

WHAT’S NEW?: Evidence of efficacy has grown

This is the first RCT to evaluate biliopancreatic diversion and only the second to evaluate gastric bypass as strategies for controlling diabetes. Similar findings were demonstrated at 12 months in an RCT of 150 obese patients with diabetes in which intensive medical therapy was compared with either gastric bypass or sleeve gastrectomy,⁶ published concurrently with the Mingrone study. Like the Mingrone study, this study found that for select patients with diabetes, surgery may lead to better outcomes than medical management alone.

CAVEATS: Long-term effect is still uncertain

The long-term efficacy of surgery as a way to manage diabetes remains uncertain. Patients in this study were followed for just 2 years and the outcomes were metabolic measures rather than morbidity and mortality. A recent prospective observational study following patients for 6 years after gastric bypass found that the rate of remission for diabetes was 75% (95% confidence interval (CI), 63%-87%) at 2 years but dropped to 62% (95% CI, 49%-75%) at 6 years⁷

A larger study (N=4047) of longer duration—the Swedish Obese Subjects (SOS) cohort study —found a considerably larger drop: The diabetes remission rate for those who had surgery went from 72% at 2 years to 36% at 10-year follow-up, but that was still higher than the 10-year remission rate (13%) for the matched controls.⁴ It is still not clear exactly how long diabetic remission lasts after bariatric surgery or what effect a 10-year respite from the disease will have on the long-term morbidity and mortality of patients with diabetes.

Surgical risks. In small studies such as the one by Mingrone et al,¹ it can be difficult to see the full extent of surgical complications. The much larger SOS study found low mortality rates (0.25%). But 13% of those who underwent bariatric surgery had postoperative complications (number needed to harm = 8), with 2.2% of patients requiring reoperation.⁴ Additionally, women who become pregnant after bariatric surgery are at increased risk for internal hernias or bowel obstruction during pregnancy.⁸

Furthermore, malabsorptive-type surgeries are known to cause nutritional deficiencies, leading to disorders including anemia and osteoporosis.⁶ Importantly, while women of childbearing-age who undergo bariatric surgery decrease their risk of developing gestational hypertension and gestational diabetes, they are more likely to have nutritional deficiencies during pregnancy and to have children with these deficiencies.⁸

CHALLENGES TO IMPLEMENTATION: The ideal candidate remains unclear

It is still not clear from this study which patients should be referred for bariatric surgery. Historically, BMI has been used as the main indication for bariatric surgery, but this and other, studies have found that remission of diabetes is independent of BMI and the amount of weight lost.⁹ A predictive 10-point Diabetes Surgery Score has recently been developed: It uses age, BMI, duration of diabetes, and C-peptide levels to predict the likelihood of diabetes remission after surgery.¹⁰ This scoring system has yet to be validated in non-Asian patients, and a threshold for recommending surgery has been not established. However, this tool indicates that younger patients with a shorter duration of diabetes (which was not a factor in the outcome of the Mingrone study) and no baseline use of insulin are most likely to benefit from surgery. Thus, these patients may be the ones we need to consider referring first.

Cost of surgery. Several studies have shown that bariatric surgery is cost-effective for the treatment of diabetes, and saves money after approximately 5 years.^11,12 However, patients with diabetes and obesity may be uninsured or underinsured, and have high out-of-pocket costs. One challenge will be to ensure that surgery is a viable option for patients with financial constraints.

Acknowledgement

The PURLs Surveillance System was supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Practice Changer
Treat low-risk patients with pulmonary embolism (PE) with low-molecular-weight heparin (LMWH) in an outpatient setting.¹

Strength of recommendation
B: Based on one good quality randomized controlled trial (RCT).

ILLUSTRATIVE CASE
Three months after undergoing surgical repair of an ankle fracture, a 50-year-old woman presents with acute-onset dyspnea at rest and pleuritic chest pain. Her left calf is tender and swollen. The patient has a history of hypertension and smokes about 10 cigarettes per day. Her temperature is 37ºC; ventricular rate, 98 beats/min; blood pressure, 135/85 mm Hg; respiratory rate, 25 breaths/min; and pulse oximetry, 92%. Spiral CT reveals a contrast filling defect indicative of a PE. Her score on the Pulmonary Embolism Severity Index (PESI) is 50, an indication of low risk. She wants to know if she can be treated at home. What should you tell her?

In the past, intravenous unfractionated heparin, administered in an inpatient setting, was the recommended initial anticoagulation therapy for patients with venous thromboembolism (VTE). LMWH, which can be administered subcutaneously and does not require laboratory monitoring, has made it possible to treat VTE without

hospitalization.

Outpatient PE care hindered

by lack of evidence
Guidelines from the American College of Physicians, the American Academy of Family Physicians, and the British Thoracic Society recommend outpatient treatment of deep vein thrombosis with LMWH, which they find to be safe and cost-effective for select patients.^2,3 Until recently, the safety and efficacy of outpatient management of PE has been less clear.

The lack of an accurate prediction tool to identify patients who could be treated safely outside the hospital was one barrier to the development of evidence-based recommendations for outpatient PE treatment. In 2005, the PESI,⁴ a validated tool that identifies patients with low risk for death from PE, was developed. Until recently, the absence of an RCT comparing inpatient and outpatient treatment for acute PE was another barrier.

STUDY SUMMARY
Outpatient treatment measures up
The Outpatient Treatment of Pulmonary Embolism (OTPE) study compared outpatient vs inpatient treatment of low-risk patients with acute PE. Participants had to be 18 or older, have acute symptomatic and objectively verified PE, and be at low risk for death based on the PESI score.4 In addition to excluding patients at moderate or high risk, the researchers identified 14 other exclusion criteria, including hypoxia, chest pain requiring opiates, and high risk for bleeding.

Patients were randomly assigned to the outpatient (n = 171) or inpatient (n = 168) group. Both groups received subcutaneous LMWH (enoxaparin, 1 mg/kg bid) for ≥ 5 days, followed by oral anticoagulation with a vitamin K antagonist for ≥ 90 days. Patients in the outpatient group were discharged from the emergency department (ED) within

24 hours of randomization, after being trained by a nurse to self-inject. Therapy after discharge was managed either by the patient's primary care physician or the hospital's anticoagulation staff.

The LMWH was discontinued in patients with an INR ≥ 2.0 for two consecutive days. All patients were followed for 90 days and contacted by the study team daily for the first week and then at 14, 30, 60, and 90 days. On each occasion, participants were asked about symptoms of recurrent VTE, bleeding, and the use of health care resources.

The primary outcome was the recurrence of symptomatic, objectively confirmed VTE within the study period. Secondary outcomes were major bleeding and all-cause mortality. Outcomes were confirmed by clinicians who were unaware of treatment

assignments.

Patients were also asked to rate their overall satisfaction with their care and their treatment preference 14 days after randomization, using a 5-point Likert questionnaire. Prior to the trial, the investigators decided that outpatient treatment would be considered noninferior to inpatient care if the difference between rates of recurrent VTE did not exceed 4%, a measure used in previous studies comparing treatment regimens for VTE and outpatient versus inpatient treatment of DVT.^5,6

Little difference in readmission rates, ED or office visits
One in 171 outpatients (0.6%) and none of the inpatients had recurrent VTE. Two outpatients (1.2%)—and no inpatients—developed major bleeding within 14 days, the result of intramuscular hematomas that occurred on days 3 and 13. There was one additional bleeding event (menometrorrhagia) in the outpatient group on day 50, but it was believed to be unrelated to the PE treatment. Per-protocol analysis, a more conservative measure used in noninferiority studies, found a difference in major bleeding rates of 3.8%. One person in each group died of non-VTE and nontreatment-related causes.

Almost all participants (99%) completed the satisfaction survey, which indicated that 92% of outpatients and 95% of inpatients were satisfied or very satisfied with their care. Hospital readmission rates, ED visits, and visits to primary care physicians were similar, with no significant differences between the

two groups. The mean time spent in the hospital was 0.5 days for outpatients and 3.9 days for inpatients. Fourteen percent of outpatients and 6% of inpatients received home nursing visits for enoxaparin injection. The total number of home visits was higher among outpatients (348 vs 105). Because both groups had extreme outliers, however, this difference was not statistically significant.

WHAT'S NEW
It's safe to keep

low-risk patients at home
This is the first RCT comparing the safety and effectiveness of outpatient and inpatient treatment of acute, symptomatic PE. Results were statistically comparable, and patients were satisfied being treated at home. Outpatient treatment was less expensive because of the shorter length of stay (0.5 vs 3.9 days) and was associated with the same rates of hospital readmission, ED visits, and visits to primary care physicians. There were more home nursing visits in the outpatient treatment group. But even if you assume a cost of $200 per home visit, the additional cost would be about $282 per individual in the outpatient group—significantly less than the cost of the additional 3.4 days in the hospital for each individual in the inpatient group.

The study also confirmed that the PESI accurately identifies low-risk patients with PE who can be treated in an outpatient setting. Thirty percent of patients who were screened for the OTPE trial met the low-risk eligibility requirement.

CAVEATS
Use of risk assessment tool is essential
The average age of patients in this study was 47 in the outpatient group and 49 in the inpatient group. In addition, only 1% to 3% of the patients were diagnosed with cancer. Older patients who have both cancer and PE would be unlikely to qualify for outpatient care.

Clinicians applying this practice changer should use the PESI to ensure that outpatient treatment for PE is used only for individuals at low risk.

CHALLENGES TO IMPLEMENTATION
ED coordination, training, and home care won't be easy
This practice changer may be difficult for primary care providers, who might not be included in emergency physicians' decisions regarding the appropriate treatment for acute PE. In this study, primary care physicians were notified of the randomized treatment plan for their patients, and 17 potential participants were excluded from the trial because of their doctors' opposition.

Outpatient management should be considered only if arrangements for adequate home nursing care can be made, if needed—and only for patients who are able to follow instructions and self-inject LMWH. Newer anticoagulation medications that are either injected once a day or taken orally might decrease the need for home nursing visits.  

REFERENCES
1. Aujesky D, Roy PM, Verschuren F, et al. Outpatient versus inpatient treatment for patients with acute pulmonary embolism: an international, open-label, randomised, non-inferiority trial. Lancet. 2011; 378:41-48.

2. Snow V, Qaseem A, Barry P, et al. Management of venous thromboembolism: a clinical practice guideline from the American College of Physicians and the American Academy of Family Physicians. Ann Intern Med. 2007;146:204-210.

3. British Thoracic Society Standards of Care Committee Pulmonary Embolism Guideline Development Group. British Thoracic Society guidelines for the management of suspected acute pulmonary embolism. Thorax. 2003;58: 470-483.

4.  Aujesky D, Obrosky DS, Stone RA, et al. Derivation and validation of a prognostic model for pulmonary embolism. Am J Respir Crit Care Med. 2005;172:1041-1046.

5. Koopman MM, Prandoni P, Piovella F, et al. Treatment of venous thrombosis with intravenous unfractionated heparin administered in the hospital as compared with subcutaneous low-molecular-weight heparin administered at home. The Tasman Study Group. N Engl J Med. 1996;334: 682-687.

6. Schulman S, Kearon C, Kakkar AK, et al. Dabigatran versus warfarin in the treatment of acute venous thromboembolism. N Engl J Med. 2009; 361:2342-2352.

ACKNOWLEDGEMENT
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Copyright © 2012 The Family Physicians Inquiries Network. All rights reserved.

Reprinted with permission from the Family Physicians Inquiries Network and The Journal of Family Practice. 2012;61(6):349-352.

Practice Changer
Treat low-risk patients with pulmonary embolism (PE) with low-molecular-weight heparin (LMWH) in an outpatient setting.¹

Strength of recommendation
B: Based on one good quality randomized controlled trial (RCT).

ILLUSTRATIVE CASE
Three months after undergoing surgical repair of an ankle fracture, a 50-year-old woman presents with acute-onset dyspnea at rest and pleuritic chest pain. Her left calf is tender and swollen. The patient has a history of hypertension and smokes about 10 cigarettes per day. Her temperature is 37ºC; ventricular rate, 98 beats/min; blood pressure, 135/85 mm Hg; respiratory rate, 25 breaths/min; and pulse oximetry, 92%. Spiral CT reveals a contrast filling defect indicative of a PE. Her score on the Pulmonary Embolism Severity Index (PESI) is 50, an indication of low risk. She wants to know if she can be treated at home. What should you tell her?

In the past, intravenous unfractionated heparin, administered in an inpatient setting, was the recommended initial anticoagulation therapy for patients with venous thromboembolism (VTE). LMWH, which can be administered subcutaneously and does not require laboratory monitoring, has made it possible to treat VTE without

hospitalization.

Outpatient PE care hindered

by lack of evidence
Guidelines from the American College of Physicians, the American Academy of Family Physicians, and the British Thoracic Society recommend outpatient treatment of deep vein thrombosis with LMWH, which they find to be safe and cost-effective for select patients.^2,3 Until recently, the safety and efficacy of outpatient management of PE has been less clear.

The lack of an accurate prediction tool to identify patients who could be treated safely outside the hospital was one barrier to the development of evidence-based recommendations for outpatient PE treatment. In 2005, the PESI,⁴ a validated tool that identifies patients with low risk for death from PE, was developed. Until recently, the absence of an RCT comparing inpatient and outpatient treatment for acute PE was another barrier.

STUDY SUMMARY
Outpatient treatment measures up
The Outpatient Treatment of Pulmonary Embolism (OTPE) study compared outpatient vs inpatient treatment of low-risk patients with acute PE. Participants had to be 18 or older, have acute symptomatic and objectively verified PE, and be at low risk for death based on the PESI score.4 In addition to excluding patients at moderate or high risk, the researchers identified 14 other exclusion criteria, including hypoxia, chest pain requiring opiates, and high risk for bleeding.

Patients were randomly assigned to the outpatient (n = 171) or inpatient (n = 168) group. Both groups received subcutaneous LMWH (enoxaparin, 1 mg/kg bid) for ≥ 5 days, followed by oral anticoagulation with a vitamin K antagonist for ≥ 90 days. Patients in the outpatient group were discharged from the emergency department (ED) within

24 hours of randomization, after being trained by a nurse to self-inject. Therapy after discharge was managed either by the patient's primary care physician or the hospital's anticoagulation staff.

The LMWH was discontinued in patients with an INR ≥ 2.0 for two consecutive days. All patients were followed for 90 days and contacted by the study team daily for the first week and then at 14, 30, 60, and 90 days. On each occasion, participants were asked about symptoms of recurrent VTE, bleeding, and the use of health care resources.

The primary outcome was the recurrence of symptomatic, objectively confirmed VTE within the study period. Secondary outcomes were major bleeding and all-cause mortality. Outcomes were confirmed by clinicians who were unaware of treatment

assignments.

Patients were also asked to rate their overall satisfaction with their care and their treatment preference 14 days after randomization, using a 5-point Likert questionnaire. Prior to the trial, the investigators decided that outpatient treatment would be considered noninferior to inpatient care if the difference between rates of recurrent VTE did not exceed 4%, a measure used in previous studies comparing treatment regimens for VTE and outpatient versus inpatient treatment of DVT.^5,6

Little difference in readmission rates, ED or office visits
One in 171 outpatients (0.6%) and none of the inpatients had recurrent VTE. Two outpatients (1.2%)—and no inpatients—developed major bleeding within 14 days, the result of intramuscular hematomas that occurred on days 3 and 13. There was one additional bleeding event (menometrorrhagia) in the outpatient group on day 50, but it was believed to be unrelated to the PE treatment. Per-protocol analysis, a more conservative measure used in noninferiority studies, found a difference in major bleeding rates of 3.8%. One person in each group died of non-VTE and nontreatment-related causes.

Almost all participants (99%) completed the satisfaction survey, which indicated that 92% of outpatients and 95% of inpatients were satisfied or very satisfied with their care. Hospital readmission rates, ED visits, and visits to primary care physicians were similar, with no significant differences between the

two groups. The mean time spent in the hospital was 0.5 days for outpatients and 3.9 days for inpatients. Fourteen percent of outpatients and 6% of inpatients received home nursing visits for enoxaparin injection. The total number of home visits was higher among outpatients (348 vs 105). Because both groups had extreme outliers, however, this difference was not statistically significant.

WHAT'S NEW
It's safe to keep

low-risk patients at home
This is the first RCT comparing the safety and effectiveness of outpatient and inpatient treatment of acute, symptomatic PE. Results were statistically comparable, and patients were satisfied being treated at home. Outpatient treatment was less expensive because of the shorter length of stay (0.5 vs 3.9 days) and was associated with the same rates of hospital readmission, ED visits, and visits to primary care physicians. There were more home nursing visits in the outpatient treatment group. But even if you assume a cost of $200 per home visit, the additional cost would be about $282 per individual in the outpatient group—significantly less than the cost of the additional 3.4 days in the hospital for each individual in the inpatient group.

The study also confirmed that the PESI accurately identifies low-risk patients with PE who can be treated in an outpatient setting. Thirty percent of patients who were screened for the OTPE trial met the low-risk eligibility requirement.

CAVEATS
Use of risk assessment tool is essential
The average age of patients in this study was 47 in the outpatient group and 49 in the inpatient group. In addition, only 1% to 3% of the patients were diagnosed with cancer. Older patients who have both cancer and PE would be unlikely to qualify for outpatient care.

Clinicians applying this practice changer should use the PESI to ensure that outpatient treatment for PE is used only for individuals at low risk.

CHALLENGES TO IMPLEMENTATION
ED coordination, training, and home care won't be easy
This practice changer may be difficult for primary care providers, who might not be included in emergency physicians' decisions regarding the appropriate treatment for acute PE. In this study, primary care physicians were notified of the randomized treatment plan for their patients, and 17 potential participants were excluded from the trial because of their doctors' opposition.

Outpatient management should be considered only if arrangements for adequate home nursing care can be made, if needed—and only for patients who are able to follow instructions and self-inject LMWH. Newer anticoagulation medications that are either injected once a day or taken orally might decrease the need for home nursing visits.  

REFERENCES
1. Aujesky D, Roy PM, Verschuren F, et al. Outpatient versus inpatient treatment for patients with acute pulmonary embolism: an international, open-label, randomised, non-inferiority trial. Lancet. 2011; 378:41-48.

2. Snow V, Qaseem A, Barry P, et al. Management of venous thromboembolism: a clinical practice guideline from the American College of Physicians and the American Academy of Family Physicians. Ann Intern Med. 2007;146:204-210.

3. British Thoracic Society Standards of Care Committee Pulmonary Embolism Guideline Development Group. British Thoracic Society guidelines for the management of suspected acute pulmonary embolism. Thorax. 2003;58: 470-483.

4.  Aujesky D, Obrosky DS, Stone RA, et al. Derivation and validation of a prognostic model for pulmonary embolism. Am J Respir Crit Care Med. 2005;172:1041-1046.

5. Koopman MM, Prandoni P, Piovella F, et al. Treatment of venous thrombosis with intravenous unfractionated heparin administered in the hospital as compared with subcutaneous low-molecular-weight heparin administered at home. The Tasman Study Group. N Engl J Med. 1996;334: 682-687.

6. Schulman S, Kearon C, Kakkar AK, et al. Dabigatran versus warfarin in the treatment of acute venous thromboembolism. N Engl J Med. 2009; 361:2342-2352.

ACKNOWLEDGEMENT
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Copyright © 2012 The Family Physicians Inquiries Network. All rights reserved.

Reprinted with permission from the Family Physicians Inquiries Network and The Journal of Family Practice. 2012;61(6):349-352.

Practice Changer
Treat low-risk patients with pulmonary embolism (PE) with low-molecular-weight heparin (LMWH) in an outpatient setting.¹

Strength of recommendation
B: Based on one good quality randomized controlled trial (RCT).

ILLUSTRATIVE CASE
Three months after undergoing surgical repair of an ankle fracture, a 50-year-old woman presents with acute-onset dyspnea at rest and pleuritic chest pain. Her left calf is tender and swollen. The patient has a history of hypertension and smokes about 10 cigarettes per day. Her temperature is 37ºC; ventricular rate, 98 beats/min; blood pressure, 135/85 mm Hg; respiratory rate, 25 breaths/min; and pulse oximetry, 92%. Spiral CT reveals a contrast filling defect indicative of a PE. Her score on the Pulmonary Embolism Severity Index (PESI) is 50, an indication of low risk. She wants to know if she can be treated at home. What should you tell her?

In the past, intravenous unfractionated heparin, administered in an inpatient setting, was the recommended initial anticoagulation therapy for patients with venous thromboembolism (VTE). LMWH, which can be administered subcutaneously and does not require laboratory monitoring, has made it possible to treat VTE without

hospitalization.

Outpatient PE care hindered

by lack of evidence
Guidelines from the American College of Physicians, the American Academy of Family Physicians, and the British Thoracic Society recommend outpatient treatment of deep vein thrombosis with LMWH, which they find to be safe and cost-effective for select patients.^2,3 Until recently, the safety and efficacy of outpatient management of PE has been less clear.

The lack of an accurate prediction tool to identify patients who could be treated safely outside the hospital was one barrier to the development of evidence-based recommendations for outpatient PE treatment. In 2005, the PESI,⁴ a validated tool that identifies patients with low risk for death from PE, was developed. Until recently, the absence of an RCT comparing inpatient and outpatient treatment for acute PE was another barrier.

STUDY SUMMARY
Outpatient treatment measures up
The Outpatient Treatment of Pulmonary Embolism (OTPE) study compared outpatient vs inpatient treatment of low-risk patients with acute PE. Participants had to be 18 or older, have acute symptomatic and objectively verified PE, and be at low risk for death based on the PESI score.4 In addition to excluding patients at moderate or high risk, the researchers identified 14 other exclusion criteria, including hypoxia, chest pain requiring opiates, and high risk for bleeding.

Patients were randomly assigned to the outpatient (n = 171) or inpatient (n = 168) group. Both groups received subcutaneous LMWH (enoxaparin, 1 mg/kg bid) for ≥ 5 days, followed by oral anticoagulation with a vitamin K antagonist for ≥ 90 days. Patients in the outpatient group were discharged from the emergency department (ED) within

24 hours of randomization, after being trained by a nurse to self-inject. Therapy after discharge was managed either by the patient's primary care physician or the hospital's anticoagulation staff.

The LMWH was discontinued in patients with an INR ≥ 2.0 for two consecutive days. All patients were followed for 90 days and contacted by the study team daily for the first week and then at 14, 30, 60, and 90 days. On each occasion, participants were asked about symptoms of recurrent VTE, bleeding, and the use of health care resources.

The primary outcome was the recurrence of symptomatic, objectively confirmed VTE within the study period. Secondary outcomes were major bleeding and all-cause mortality. Outcomes were confirmed by clinicians who were unaware of treatment

assignments.

Patients were also asked to rate their overall satisfaction with their care and their treatment preference 14 days after randomization, using a 5-point Likert questionnaire. Prior to the trial, the investigators decided that outpatient treatment would be considered noninferior to inpatient care if the difference between rates of recurrent VTE did not exceed 4%, a measure used in previous studies comparing treatment regimens for VTE and outpatient versus inpatient treatment of DVT.^5,6

Little difference in readmission rates, ED or office visits
One in 171 outpatients (0.6%) and none of the inpatients had recurrent VTE. Two outpatients (1.2%)—and no inpatients—developed major bleeding within 14 days, the result of intramuscular hematomas that occurred on days 3 and 13. There was one additional bleeding event (menometrorrhagia) in the outpatient group on day 50, but it was believed to be unrelated to the PE treatment. Per-protocol analysis, a more conservative measure used in noninferiority studies, found a difference in major bleeding rates of 3.8%. One person in each group died of non-VTE and nontreatment-related causes.

Almost all participants (99%) completed the satisfaction survey, which indicated that 92% of outpatients and 95% of inpatients were satisfied or very satisfied with their care. Hospital readmission rates, ED visits, and visits to primary care physicians were similar, with no significant differences between the

two groups. The mean time spent in the hospital was 0.5 days for outpatients and 3.9 days for inpatients. Fourteen percent of outpatients and 6% of inpatients received home nursing visits for enoxaparin injection. The total number of home visits was higher among outpatients (348 vs 105). Because both groups had extreme outliers, however, this difference was not statistically significant.

WHAT'S NEW
It's safe to keep

low-risk patients at home
This is the first RCT comparing the safety and effectiveness of outpatient and inpatient treatment of acute, symptomatic PE. Results were statistically comparable, and patients were satisfied being treated at home. Outpatient treatment was less expensive because of the shorter length of stay (0.5 vs 3.9 days) and was associated with the same rates of hospital readmission, ED visits, and visits to primary care physicians. There were more home nursing visits in the outpatient treatment group. But even if you assume a cost of $200 per home visit, the additional cost would be about $282 per individual in the outpatient group—significantly less than the cost of the additional 3.4 days in the hospital for each individual in the inpatient group.

The study also confirmed that the PESI accurately identifies low-risk patients with PE who can be treated in an outpatient setting. Thirty percent of patients who were screened for the OTPE trial met the low-risk eligibility requirement.

CAVEATS
Use of risk assessment tool is essential
The average age of patients in this study was 47 in the outpatient group and 49 in the inpatient group. In addition, only 1% to 3% of the patients were diagnosed with cancer. Older patients who have both cancer and PE would be unlikely to qualify for outpatient care.

Clinicians applying this practice changer should use the PESI to ensure that outpatient treatment for PE is used only for individuals at low risk.

CHALLENGES TO IMPLEMENTATION
ED coordination, training, and home care won't be easy
This practice changer may be difficult for primary care providers, who might not be included in emergency physicians' decisions regarding the appropriate treatment for acute PE. In this study, primary care physicians were notified of the randomized treatment plan for their patients, and 17 potential participants were excluded from the trial because of their doctors' opposition.

Outpatient management should be considered only if arrangements for adequate home nursing care can be made, if needed—and only for patients who are able to follow instructions and self-inject LMWH. Newer anticoagulation medications that are either injected once a day or taken orally might decrease the need for home nursing visits.  

REFERENCES
1. Aujesky D, Roy PM, Verschuren F, et al. Outpatient versus inpatient treatment for patients with acute pulmonary embolism: an international, open-label, randomised, non-inferiority trial. Lancet. 2011; 378:41-48.

2. Snow V, Qaseem A, Barry P, et al. Management of venous thromboembolism: a clinical practice guideline from the American College of Physicians and the American Academy of Family Physicians. Ann Intern Med. 2007;146:204-210.

3. British Thoracic Society Standards of Care Committee Pulmonary Embolism Guideline Development Group. British Thoracic Society guidelines for the management of suspected acute pulmonary embolism. Thorax. 2003;58: 470-483.

4.  Aujesky D, Obrosky DS, Stone RA, et al. Derivation and validation of a prognostic model for pulmonary embolism. Am J Respir Crit Care Med. 2005;172:1041-1046.

5. Koopman MM, Prandoni P, Piovella F, et al. Treatment of venous thrombosis with intravenous unfractionated heparin administered in the hospital as compared with subcutaneous low-molecular-weight heparin administered at home. The Tasman Study Group. N Engl J Med. 1996;334: 682-687.

6. Schulman S, Kearon C, Kakkar AK, et al. Dabigatran versus warfarin in the treatment of acute venous thromboembolism. N Engl J Med. 2009; 361:2342-2352.

ACKNOWLEDGEMENT
The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Copyright © 2012 The Family Physicians Inquiries Network. All rights reserved.

Reprinted with permission from the Family Physicians Inquiries Network and The Journal of Family Practice. 2012;61(6):349-352.

ILLUSTRATIVE CASE

Three months after undergoing surgical repair of an ankle fracture, a 50-year-old woman presents with acute onset dyspnea at rest and pleuritic chest pain. Her left calf is tender and swollen. The patient has a history of hypertension and smokes about 10 cigarettes per day. Her temperature is 37°C (99°F); pulse rate, 98; blood pressure, 135/85 mm Hg; respiratory rate, 25; and pulse oximetry, 92%.

You order a spiral CT, which reveals a contrast filling defect indicative of a PE. Her score on the Pulmonary Embolism Severity Index (PESI) is 50, an indication of low risk. She wants to know if she can be treated at home. What should you tell her?

In the past, intravenous unfractionated heparin, administered in an inpatient setting, was the recommended initial anticoagulation therapy for patients with venous thromboembolism (VTE). LMWH, which can be administered subcutaneously and does not require laboratory monitoring, has made it possible to treat VTE without hospitalization.

Outpatient PE care hindered by lack of evidence
Guidelines from the American College of Physicians, American Academy of Family Physicians, and British Thoracic Society recommend outpatient treatment of deep vein thrombosis with LMWH, which they find to be safe and cost effective for select patients.^2,3 Until recently, the safety and efficacy of out-patient management of PE has been less clear.

The lack of an accurate prediction tool to identify patients who could be treated safely outside of the hospital was one barrier to the development of evidence-based recommendations for outpatient PE treatment. In 2005, the PESI,⁴ a validated tool that identifies patients with low risk of death from PE, was developed. Until recently, the absence of an RCT comparing inpatient and outpatient treatment for acute PE was another barrier.

STUDY SUMMARY: Outpatient treatment measures up

The Outpatient Treatment of Pulmonary Embolism (OTPE) study was a multinational, randomized, noninferiority trial comparing outpatient vs inpatient treatment of low-risk patients with acute PE. Participants had to be ≥18 years old, have acute symptomatic and objectively verified PE, and be at low risk of death based on the PESI score.⁴ In addition to excluding patients at moderate or high risk, the researchers identified 14 other exclusion criteria, including hypoxia, chest pain requiring opiates, and high risk for bleeding.

Patients were randomly assigned to the outpatient (n=171) or inpatient (n=168) group. Both groups received subcutaneous LMWH (enoxaparin, 1 mg/kg twice a day) for ≥5 days, followed by oral anticoagulation with a vitamin K antagonist for ≥90 days. Patients in the outpatient group were discharged from the emergency department (ED) within 24 hours of randomization, after being trained by a nurse to self-inject. Therapy after discharge was managed either by the patient’s primary care physician or the hospital’s anticoagulation staff.

The LMWH was discontinued in patients with an INR ≥2.0 for 2 consecutive days. All patients were followed for 90 days, and contacted by the study team daily for the first week and then at 14, 30, 60, and 90 days. On each occasion, participants were asked about symptoms of recurrent VTE, bleeding, and the use of health care resources.

The primary outcome was the recurrence of symptomatic, objectively confirmed VTE within the study period. Secondary outcomes were major bleeding and all-cause mortality. Outcomes were confirmed by clinicians who were unaware of treatment assignments.

Patients were also asked to rate both their overall satisfaction with their care and their treatment preference 14 days after randomization, using a 5-point Likert questionnaire. Prior to the trial, the investigators decided that outpatient treatment would be considered noninferior to inpatient care if the difference between rates of recurrent VTE did not exceed 4%, a measure used in previous studies comparing treatment regimens for VTE and outpatient vs inpatient treatment of DVT.^5,6

Little difference in readmission rates, ED or office visits
One in 171 outpatients (0.6%) and none of the inpatients had recurrent VTE. Two out-patients (1.2%)—and no inpatients—developed major bleeding within 14 days, the result of intramuscular hematomas that occurred on Days 3 and 13. There was one additional bleeding event (menometrorrhagia) in the outpatient group on Day 50, but it was believed to be unrelated to the PE treatment. Per-protocol analysis, a more conservative measure used in noninferiority studies, found a difference in major bleeding rates of 3.8%. One person in each group died from non-VTE and nontreatment-related causes.

Almost all participants (99%) completed the satisfaction survey, which indicated that 92% of outpatients and 95% of inpatients were satisfied or very satisfied with their care. Hospital readmission rates, ED visits, and visits to primary care physicians were similar, with no significant differences between the 2 groups. The mean time spent in the hospital was 0.5 days (standard deviation [SD], 1.0) for outpatients and 3.9 days (SD, 3.1) for in-patients. Fourteen percent of outpatients and 6% of inpatients received home nursing visits for enoxaparin injection. The total number of home visits was higher among outpatients (348 vs 105). Because both groups had extreme outliers, however, this difference was not statistically significant.

WHAT’S NEW: It’s safe to keep low-risk patients at home

This is the first RCT comparing the safety and effectiveness of outpatient and inpatient treatment of acute, symptomatic PE. Results were statistically comparable, and patients were satisfied being treated at home. Outpatient treatment was less expensive because of the shorter length of stay (0.5 vs 3.9 days) and was associated with the same rates of hospital readmission, ED visits, and visits to primary care physicians. There were more home nursing visits in the outpatient treatment group. But even if you assume a cost of $200 per home visit, the additional cost would be about $282 per individual in the outpatient group—significantly less than the cost of the additional 3.4 days in the hospital for each individual in the inpatient group.

The study also confirmed that the PESI accurately identifies low-risk patients with PE who can be treated in an outpatient setting. Thirty percent of patients who were screened for the OTPE trial met the low-risk eligibility requirement.

CAVEATS: Use of risk assessment tool is essential

The average age of patients in this study was 47 years in the outpatient group and 49 years in the inpatient group. In addition, only 1% to 3% of the patients were diagnosed with cancer. Older patients who have both cancer and PE would be unlikely to qualify for outpatient care.

Physicians applying this practice changer should use the PESI to ensure that outpatient treatment for PE is used only for individuals at low risk.

CHALLENGES TO IMPLEMENTATION: ED coordination, training, and home care won’t be easy

This practice changer may be difficult for family physicians, who might not be included in emergency physicians’ decisions regarding the appropriate treatment for acute PE. In this study, primary care physicians were notified of the randomized treatment plan for their patients, and 17 potential participants were excluded from the trial because of their doctors’ opposition.

Outpatient management should be considered only if arrangements for adequate home nursing care can be made, if needed— and only for patients who are able to follow instructions and self-inject LMWH. Newer anticoagulation medications that are either injected once a day or taken orally might decrease the need for home nursing visits.

Acknowledgement

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

Three months after undergoing surgical repair of an ankle fracture, a 50-year-old woman presents with acute onset dyspnea at rest and pleuritic chest pain. Her left calf is tender and swollen. The patient has a history of hypertension and smokes about 10 cigarettes per day. Her temperature is 37°C (99°F); pulse rate, 98; blood pressure, 135/85 mm Hg; respiratory rate, 25; and pulse oximetry, 92%.

You order a spiral CT, which reveals a contrast filling defect indicative of a PE. Her score on the Pulmonary Embolism Severity Index (PESI) is 50, an indication of low risk. She wants to know if she can be treated at home. What should you tell her?

In the past, intravenous unfractionated heparin, administered in an inpatient setting, was the recommended initial anticoagulation therapy for patients with venous thromboembolism (VTE). LMWH, which can be administered subcutaneously and does not require laboratory monitoring, has made it possible to treat VTE without hospitalization.

Outpatient PE care hindered by lack of evidence
Guidelines from the American College of Physicians, American Academy of Family Physicians, and British Thoracic Society recommend outpatient treatment of deep vein thrombosis with LMWH, which they find to be safe and cost effective for select patients.^2,3 Until recently, the safety and efficacy of out-patient management of PE has been less clear.

The lack of an accurate prediction tool to identify patients who could be treated safely outside of the hospital was one barrier to the development of evidence-based recommendations for outpatient PE treatment. In 2005, the PESI,⁴ a validated tool that identifies patients with low risk of death from PE, was developed. Until recently, the absence of an RCT comparing inpatient and outpatient treatment for acute PE was another barrier.

STUDY SUMMARY: Outpatient treatment measures up

The Outpatient Treatment of Pulmonary Embolism (OTPE) study was a multinational, randomized, noninferiority trial comparing outpatient vs inpatient treatment of low-risk patients with acute PE. Participants had to be ≥18 years old, have acute symptomatic and objectively verified PE, and be at low risk of death based on the PESI score.⁴ In addition to excluding patients at moderate or high risk, the researchers identified 14 other exclusion criteria, including hypoxia, chest pain requiring opiates, and high risk for bleeding.

Patients were randomly assigned to the outpatient (n=171) or inpatient (n=168) group. Both groups received subcutaneous LMWH (enoxaparin, 1 mg/kg twice a day) for ≥5 days, followed by oral anticoagulation with a vitamin K antagonist for ≥90 days. Patients in the outpatient group were discharged from the emergency department (ED) within 24 hours of randomization, after being trained by a nurse to self-inject. Therapy after discharge was managed either by the patient’s primary care physician or the hospital’s anticoagulation staff.

The LMWH was discontinued in patients with an INR ≥2.0 for 2 consecutive days. All patients were followed for 90 days, and contacted by the study team daily for the first week and then at 14, 30, 60, and 90 days. On each occasion, participants were asked about symptoms of recurrent VTE, bleeding, and the use of health care resources.

The primary outcome was the recurrence of symptomatic, objectively confirmed VTE within the study period. Secondary outcomes were major bleeding and all-cause mortality. Outcomes were confirmed by clinicians who were unaware of treatment assignments.

Patients were also asked to rate both their overall satisfaction with their care and their treatment preference 14 days after randomization, using a 5-point Likert questionnaire. Prior to the trial, the investigators decided that outpatient treatment would be considered noninferior to inpatient care if the difference between rates of recurrent VTE did not exceed 4%, a measure used in previous studies comparing treatment regimens for VTE and outpatient vs inpatient treatment of DVT.^5,6

Little difference in readmission rates, ED or office visits
One in 171 outpatients (0.6%) and none of the inpatients had recurrent VTE. Two out-patients (1.2%)—and no inpatients—developed major bleeding within 14 days, the result of intramuscular hematomas that occurred on Days 3 and 13. There was one additional bleeding event (menometrorrhagia) in the outpatient group on Day 50, but it was believed to be unrelated to the PE treatment. Per-protocol analysis, a more conservative measure used in noninferiority studies, found a difference in major bleeding rates of 3.8%. One person in each group died from non-VTE and nontreatment-related causes.

Almost all participants (99%) completed the satisfaction survey, which indicated that 92% of outpatients and 95% of inpatients were satisfied or very satisfied with their care. Hospital readmission rates, ED visits, and visits to primary care physicians were similar, with no significant differences between the 2 groups. The mean time spent in the hospital was 0.5 days (standard deviation [SD], 1.0) for outpatients and 3.9 days (SD, 3.1) for in-patients. Fourteen percent of outpatients and 6% of inpatients received home nursing visits for enoxaparin injection. The total number of home visits was higher among outpatients (348 vs 105). Because both groups had extreme outliers, however, this difference was not statistically significant.

WHAT’S NEW: It’s safe to keep low-risk patients at home

This is the first RCT comparing the safety and effectiveness of outpatient and inpatient treatment of acute, symptomatic PE. Results were statistically comparable, and patients were satisfied being treated at home. Outpatient treatment was less expensive because of the shorter length of stay (0.5 vs 3.9 days) and was associated with the same rates of hospital readmission, ED visits, and visits to primary care physicians. There were more home nursing visits in the outpatient treatment group. But even if you assume a cost of $200 per home visit, the additional cost would be about $282 per individual in the outpatient group—significantly less than the cost of the additional 3.4 days in the hospital for each individual in the inpatient group.

The study also confirmed that the PESI accurately identifies low-risk patients with PE who can be treated in an outpatient setting. Thirty percent of patients who were screened for the OTPE trial met the low-risk eligibility requirement.

CAVEATS: Use of risk assessment tool is essential

The average age of patients in this study was 47 years in the outpatient group and 49 years in the inpatient group. In addition, only 1% to 3% of the patients were diagnosed with cancer. Older patients who have both cancer and PE would be unlikely to qualify for outpatient care.

Physicians applying this practice changer should use the PESI to ensure that outpatient treatment for PE is used only for individuals at low risk.

CHALLENGES TO IMPLEMENTATION: ED coordination, training, and home care won’t be easy

This practice changer may be difficult for family physicians, who might not be included in emergency physicians’ decisions regarding the appropriate treatment for acute PE. In this study, primary care physicians were notified of the randomized treatment plan for their patients, and 17 potential participants were excluded from the trial because of their doctors’ opposition.

Outpatient management should be considered only if arrangements for adequate home nursing care can be made, if needed— and only for patients who are able to follow instructions and self-inject LMWH. Newer anticoagulation medications that are either injected once a day or taken orally might decrease the need for home nursing visits.

Acknowledgement

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

Three months after undergoing surgical repair of an ankle fracture, a 50-year-old woman presents with acute onset dyspnea at rest and pleuritic chest pain. Her left calf is tender and swollen. The patient has a history of hypertension and smokes about 10 cigarettes per day. Her temperature is 37°C (99°F); pulse rate, 98; blood pressure, 135/85 mm Hg; respiratory rate, 25; and pulse oximetry, 92%.

You order a spiral CT, which reveals a contrast filling defect indicative of a PE. Her score on the Pulmonary Embolism Severity Index (PESI) is 50, an indication of low risk. She wants to know if she can be treated at home. What should you tell her?

In the past, intravenous unfractionated heparin, administered in an inpatient setting, was the recommended initial anticoagulation therapy for patients with venous thromboembolism (VTE). LMWH, which can be administered subcutaneously and does not require laboratory monitoring, has made it possible to treat VTE without hospitalization.

Outpatient PE care hindered by lack of evidence
Guidelines from the American College of Physicians, American Academy of Family Physicians, and British Thoracic Society recommend outpatient treatment of deep vein thrombosis with LMWH, which they find to be safe and cost effective for select patients.^2,3 Until recently, the safety and efficacy of out-patient management of PE has been less clear.

The lack of an accurate prediction tool to identify patients who could be treated safely outside of the hospital was one barrier to the development of evidence-based recommendations for outpatient PE treatment. In 2005, the PESI,⁴ a validated tool that identifies patients with low risk of death from PE, was developed. Until recently, the absence of an RCT comparing inpatient and outpatient treatment for acute PE was another barrier.

STUDY SUMMARY: Outpatient treatment measures up

The Outpatient Treatment of Pulmonary Embolism (OTPE) study was a multinational, randomized, noninferiority trial comparing outpatient vs inpatient treatment of low-risk patients with acute PE. Participants had to be ≥18 years old, have acute symptomatic and objectively verified PE, and be at low risk of death based on the PESI score.⁴ In addition to excluding patients at moderate or high risk, the researchers identified 14 other exclusion criteria, including hypoxia, chest pain requiring opiates, and high risk for bleeding.

Patients were randomly assigned to the outpatient (n=171) or inpatient (n=168) group. Both groups received subcutaneous LMWH (enoxaparin, 1 mg/kg twice a day) for ≥5 days, followed by oral anticoagulation with a vitamin K antagonist for ≥90 days. Patients in the outpatient group were discharged from the emergency department (ED) within 24 hours of randomization, after being trained by a nurse to self-inject. Therapy after discharge was managed either by the patient’s primary care physician or the hospital’s anticoagulation staff.

The LMWH was discontinued in patients with an INR ≥2.0 for 2 consecutive days. All patients were followed for 90 days, and contacted by the study team daily for the first week and then at 14, 30, 60, and 90 days. On each occasion, participants were asked about symptoms of recurrent VTE, bleeding, and the use of health care resources.

The primary outcome was the recurrence of symptomatic, objectively confirmed VTE within the study period. Secondary outcomes were major bleeding and all-cause mortality. Outcomes were confirmed by clinicians who were unaware of treatment assignments.

Patients were also asked to rate both their overall satisfaction with their care and their treatment preference 14 days after randomization, using a 5-point Likert questionnaire. Prior to the trial, the investigators decided that outpatient treatment would be considered noninferior to inpatient care if the difference between rates of recurrent VTE did not exceed 4%, a measure used in previous studies comparing treatment regimens for VTE and outpatient vs inpatient treatment of DVT.^5,6

Little difference in readmission rates, ED or office visits
One in 171 outpatients (0.6%) and none of the inpatients had recurrent VTE. Two out-patients (1.2%)—and no inpatients—developed major bleeding within 14 days, the result of intramuscular hematomas that occurred on Days 3 and 13. There was one additional bleeding event (menometrorrhagia) in the outpatient group on Day 50, but it was believed to be unrelated to the PE treatment. Per-protocol analysis, a more conservative measure used in noninferiority studies, found a difference in major bleeding rates of 3.8%. One person in each group died from non-VTE and nontreatment-related causes.

Almost all participants (99%) completed the satisfaction survey, which indicated that 92% of outpatients and 95% of inpatients were satisfied or very satisfied with their care. Hospital readmission rates, ED visits, and visits to primary care physicians were similar, with no significant differences between the 2 groups. The mean time spent in the hospital was 0.5 days (standard deviation [SD], 1.0) for outpatients and 3.9 days (SD, 3.1) for in-patients. Fourteen percent of outpatients and 6% of inpatients received home nursing visits for enoxaparin injection. The total number of home visits was higher among outpatients (348 vs 105). Because both groups had extreme outliers, however, this difference was not statistically significant.

WHAT’S NEW: It’s safe to keep low-risk patients at home

This is the first RCT comparing the safety and effectiveness of outpatient and inpatient treatment of acute, symptomatic PE. Results were statistically comparable, and patients were satisfied being treated at home. Outpatient treatment was less expensive because of the shorter length of stay (0.5 vs 3.9 days) and was associated with the same rates of hospital readmission, ED visits, and visits to primary care physicians. There were more home nursing visits in the outpatient treatment group. But even if you assume a cost of $200 per home visit, the additional cost would be about $282 per individual in the outpatient group—significantly less than the cost of the additional 3.4 days in the hospital for each individual in the inpatient group.

The study also confirmed that the PESI accurately identifies low-risk patients with PE who can be treated in an outpatient setting. Thirty percent of patients who were screened for the OTPE trial met the low-risk eligibility requirement.

CAVEATS: Use of risk assessment tool is essential

The average age of patients in this study was 47 years in the outpatient group and 49 years in the inpatient group. In addition, only 1% to 3% of the patients were diagnosed with cancer. Older patients who have both cancer and PE would be unlikely to qualify for outpatient care.

Physicians applying this practice changer should use the PESI to ensure that outpatient treatment for PE is used only for individuals at low risk.

CHALLENGES TO IMPLEMENTATION: ED coordination, training, and home care won’t be easy

This practice changer may be difficult for family physicians, who might not be included in emergency physicians’ decisions regarding the appropriate treatment for acute PE. In this study, primary care physicians were notified of the randomized treatment plan for their patients, and 17 potential participants were excluded from the trial because of their doctors’ opposition.

Outpatient management should be considered only if arrangements for adequate home nursing care can be made, if needed— and only for patients who are able to follow instructions and self-inject LMWH. Newer anticoagulation medications that are either injected once a day or taken orally might decrease the need for home nursing visits.

Acknowledgement

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

In the last trimester of pregnancy, a patient asks about her options for postpartum contraception. She plans to breastfeed and does not want to have another child for several years, she says. Her family is scheduled to move 2 weeks after her due date, and she wants to begin using contraception before then. She’s interested in the etonogestrel implant (Implanon) and wonders whether she can have it inserted before she leaves the hospital. What can you tell her?

Approximately 4 million women give birth each year in the United States,² 77% of whom choose to breastfeed their babies.³ Postpartum contraception is recommended, to ensure adequate spacing between (or prevention of) pregnancies.

Hormonal options are limited for nursing moms
Due to the negative effect of estrogens on lactation,⁴ women who wish to use birth control while breastfeeding have limited choices. Their options include progestin-only oral contraceptives; intrauterine devices, including the levonorgestrel intrauterine contraceptive; barrier methods; and the etonogestrel implant. Yet concerns remain that using a progestin contraceptive in the early postpartum period could negatively affect lactogenesis, as well as the quantity and quality of the breast milk.⁵

Starting contraception after 6 weeks? The opportunity is often missed
A 2010 systematic review found that progestin-only contraception can be safely used in breastfeeding women. However, the studies included in the review did not consider timing. Thus, the researchers concluded only that initiation of a progestin contraception >6 weeks postpartum is safe.⁶ The World Health Organization recommends waiting >6 weeks, as well.⁷ But studies have found that between 10% and 40% of women miss their 6-week postpartum visit,⁸ thereby missing the opportunity to start contraception.

A 2009 pilot study found that the implant can be safely used <4 weeks postpartum, and did not affect breastfeeding.⁹ The study we review here is the first RCT to evaluate the impact of early insertion (1-3 days postpartum) of the etonogestrel implant on lactogenesis.

STUDY SUMMARY: Timing of implant did not affect outcomes

The study by Gurtcheff et al was a randomized controlled noninferiority trial of 69 women who wanted to use Implanon for postpartum birth control. Inclusion criteria included good health (of the babies as well as the mothers), the intention to breastfeed, and the willingness to be randomly assigned to either early (1-3 days) or standard (4-8 weeks) insertion. The study was not blinded. No other source of bias was identified.

The primary outcomes studied were time to stage II of lactogenesis (based on maternal perception of when her milk “had come in”) and rates of lactation failure.

Early insertion, the researchers found, was noninferior to standard insertion, both in the time to stage II of lactogenesis and the risk of lactation failure. The time to lactogenesis was 64.3 hours (mean standard deviation [SD], 19.6 hours) for early insertion vs 65.2 hours (mean SD, 18.5 hours) for standard insertion. The mean difference was -1.4 hours (95% confidence interval [CI], -10.6 to 7.7 hours). For lactation failure, the absolute risk difference was 0.03 (95% CI, -0.02 to 0.08).

Secondary outcomes included breastfeeding status, side effects, and bleeding patterns, as well as the contraceptive method actually being used at the time. This information was gathered at 2 weeks, 6 weeks, 3 months, and 6 months postpartum.

There were no statistically significant differences in breastfeeding, formula supplementation, or patient-reported bleeding patterns. However, a third of the women (11 of 34) in the standard group did not have the implant inserted, and opted for an alternate form of birth control.

At 6 weeks, women in both groups had a milk sample analyzed for fat and energy content. There was no significant difference in mean creamatocrit values between the groups.

WHAT’S NEW: Early insertion is safe and fosters compliance

Lactogenesis and lactation failure rates were comparable, whether the etonogestrel implant was inserted between 1 and 3 days postpartum or 4 to 8 weeks postpartum. An advantage of early insertion was increased contraceptive compliance. At 3 months postpartum, 13% of the women in the standard group were not using any birth control. Among those in the early insertion group, compliance was 100%.

CAVEATS: Study sample may not be representative

This was a small study, but it was powered to detect ≥8 hour difference in onset of stage II lactogenesis. Participants were not representative of all populations (91% were white, 73% of whom were Hispanic). Both the mothers and babies were healthy, so we can’t extrapolate to situations where either mom or baby is sick.

CHALLENGES TO IMPLEMENTATION: Finding clinicians trained in insertion technique

Health care providers trained in insertion of the etonogestrel implant would need to be available to promote insertion in the early postpartum period. Ensuring availability of the device in hospitals may require extra logistical planning; incorporating etonogestrel implant insertion into already-hectic morning rounds may be challenging, as well.

Acknowledgement

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

In the last trimester of pregnancy, a patient asks about her options for postpartum contraception. She plans to breastfeed and does not want to have another child for several years, she says. Her family is scheduled to move 2 weeks after her due date, and she wants to begin using contraception before then. She’s interested in the etonogestrel implant (Implanon) and wonders whether she can have it inserted before she leaves the hospital. What can you tell her?

Approximately 4 million women give birth each year in the United States,² 77% of whom choose to breastfeed their babies.³ Postpartum contraception is recommended, to ensure adequate spacing between (or prevention of) pregnancies.

Hormonal options are limited for nursing moms
Due to the negative effect of estrogens on lactation,⁴ women who wish to use birth control while breastfeeding have limited choices. Their options include progestin-only oral contraceptives; intrauterine devices, including the levonorgestrel intrauterine contraceptive; barrier methods; and the etonogestrel implant. Yet concerns remain that using a progestin contraceptive in the early postpartum period could negatively affect lactogenesis, as well as the quantity and quality of the breast milk.⁵

Starting contraception after 6 weeks? The opportunity is often missed
A 2010 systematic review found that progestin-only contraception can be safely used in breastfeeding women. However, the studies included in the review did not consider timing. Thus, the researchers concluded only that initiation of a progestin contraception >6 weeks postpartum is safe.⁶ The World Health Organization recommends waiting >6 weeks, as well.⁷ But studies have found that between 10% and 40% of women miss their 6-week postpartum visit,⁸ thereby missing the opportunity to start contraception.

A 2009 pilot study found that the implant can be safely used <4 weeks postpartum, and did not affect breastfeeding.⁹ The study we review here is the first RCT to evaluate the impact of early insertion (1-3 days postpartum) of the etonogestrel implant on lactogenesis.

STUDY SUMMARY: Timing of implant did not affect outcomes

The study by Gurtcheff et al was a randomized controlled noninferiority trial of 69 women who wanted to use Implanon for postpartum birth control. Inclusion criteria included good health (of the babies as well as the mothers), the intention to breastfeed, and the willingness to be randomly assigned to either early (1-3 days) or standard (4-8 weeks) insertion. The study was not blinded. No other source of bias was identified.

The primary outcomes studied were time to stage II of lactogenesis (based on maternal perception of when her milk “had come in”) and rates of lactation failure.

Early insertion, the researchers found, was noninferior to standard insertion, both in the time to stage II of lactogenesis and the risk of lactation failure. The time to lactogenesis was 64.3 hours (mean standard deviation [SD], 19.6 hours) for early insertion vs 65.2 hours (mean SD, 18.5 hours) for standard insertion. The mean difference was -1.4 hours (95% confidence interval [CI], -10.6 to 7.7 hours). For lactation failure, the absolute risk difference was 0.03 (95% CI, -0.02 to 0.08).

Secondary outcomes included breastfeeding status, side effects, and bleeding patterns, as well as the contraceptive method actually being used at the time. This information was gathered at 2 weeks, 6 weeks, 3 months, and 6 months postpartum.

There were no statistically significant differences in breastfeeding, formula supplementation, or patient-reported bleeding patterns. However, a third of the women (11 of 34) in the standard group did not have the implant inserted, and opted for an alternate form of birth control.

At 6 weeks, women in both groups had a milk sample analyzed for fat and energy content. There was no significant difference in mean creamatocrit values between the groups.

WHAT’S NEW: Early insertion is safe and fosters compliance

Lactogenesis and lactation failure rates were comparable, whether the etonogestrel implant was inserted between 1 and 3 days postpartum or 4 to 8 weeks postpartum. An advantage of early insertion was increased contraceptive compliance. At 3 months postpartum, 13% of the women in the standard group were not using any birth control. Among those in the early insertion group, compliance was 100%.

CAVEATS: Study sample may not be representative

This was a small study, but it was powered to detect ≥8 hour difference in onset of stage II lactogenesis. Participants were not representative of all populations (91% were white, 73% of whom were Hispanic). Both the mothers and babies were healthy, so we can’t extrapolate to situations where either mom or baby is sick.

CHALLENGES TO IMPLEMENTATION: Finding clinicians trained in insertion technique

Health care providers trained in insertion of the etonogestrel implant would need to be available to promote insertion in the early postpartum period. Ensuring availability of the device in hospitals may require extra logistical planning; incorporating etonogestrel implant insertion into already-hectic morning rounds may be challenging, as well.

Acknowledgement

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY

ILLUSTRATIVE CASE

In the last trimester of pregnancy, a patient asks about her options for postpartum contraception. She plans to breastfeed and does not want to have another child for several years, she says. Her family is scheduled to move 2 weeks after her due date, and she wants to begin using contraception before then. She’s interested in the etonogestrel implant (Implanon) and wonders whether she can have it inserted before she leaves the hospital. What can you tell her?

Approximately 4 million women give birth each year in the United States,² 77% of whom choose to breastfeed their babies.³ Postpartum contraception is recommended, to ensure adequate spacing between (or prevention of) pregnancies.

Hormonal options are limited for nursing moms
Due to the negative effect of estrogens on lactation,⁴ women who wish to use birth control while breastfeeding have limited choices. Their options include progestin-only oral contraceptives; intrauterine devices, including the levonorgestrel intrauterine contraceptive; barrier methods; and the etonogestrel implant. Yet concerns remain that using a progestin contraceptive in the early postpartum period could negatively affect lactogenesis, as well as the quantity and quality of the breast milk.⁵

Starting contraception after 6 weeks? The opportunity is often missed
A 2010 systematic review found that progestin-only contraception can be safely used in breastfeeding women. However, the studies included in the review did not consider timing. Thus, the researchers concluded only that initiation of a progestin contraception >6 weeks postpartum is safe.⁶ The World Health Organization recommends waiting >6 weeks, as well.⁷ But studies have found that between 10% and 40% of women miss their 6-week postpartum visit,⁸ thereby missing the opportunity to start contraception.

A 2009 pilot study found that the implant can be safely used <4 weeks postpartum, and did not affect breastfeeding.⁹ The study we review here is the first RCT to evaluate the impact of early insertion (1-3 days postpartum) of the etonogestrel implant on lactogenesis.

STUDY SUMMARY: Timing of implant did not affect outcomes

The study by Gurtcheff et al was a randomized controlled noninferiority trial of 69 women who wanted to use Implanon for postpartum birth control. Inclusion criteria included good health (of the babies as well as the mothers), the intention to breastfeed, and the willingness to be randomly assigned to either early (1-3 days) or standard (4-8 weeks) insertion. The study was not blinded. No other source of bias was identified.

The primary outcomes studied were time to stage II of lactogenesis (based on maternal perception of when her milk “had come in”) and rates of lactation failure.

Early insertion, the researchers found, was noninferior to standard insertion, both in the time to stage II of lactogenesis and the risk of lactation failure. The time to lactogenesis was 64.3 hours (mean standard deviation [SD], 19.6 hours) for early insertion vs 65.2 hours (mean SD, 18.5 hours) for standard insertion. The mean difference was -1.4 hours (95% confidence interval [CI], -10.6 to 7.7 hours). For lactation failure, the absolute risk difference was 0.03 (95% CI, -0.02 to 0.08).

Secondary outcomes included breastfeeding status, side effects, and bleeding patterns, as well as the contraceptive method actually being used at the time. This information was gathered at 2 weeks, 6 weeks, 3 months, and 6 months postpartum.

There were no statistically significant differences in breastfeeding, formula supplementation, or patient-reported bleeding patterns. However, a third of the women (11 of 34) in the standard group did not have the implant inserted, and opted for an alternate form of birth control.

At 6 weeks, women in both groups had a milk sample analyzed for fat and energy content. There was no significant difference in mean creamatocrit values between the groups.

WHAT’S NEW: Early insertion is safe and fosters compliance

Lactogenesis and lactation failure rates were comparable, whether the etonogestrel implant was inserted between 1 and 3 days postpartum or 4 to 8 weeks postpartum. An advantage of early insertion was increased contraceptive compliance. At 3 months postpartum, 13% of the women in the standard group were not using any birth control. Among those in the early insertion group, compliance was 100%.

CAVEATS: Study sample may not be representative

This was a small study, but it was powered to detect ≥8 hour difference in onset of stage II lactogenesis. Participants were not representative of all populations (91% were white, 73% of whom were Hispanic). Both the mothers and babies were healthy, so we can’t extrapolate to situations where either mom or baby is sick.

CHALLENGES TO IMPLEMENTATION: Finding clinicians trained in insertion technique

Health care providers trained in insertion of the etonogestrel implant would need to be available to promote insertion in the early postpartum period. Ensuring availability of the device in hospitals may require extra logistical planning; incorporating etonogestrel implant insertion into already-hectic morning rounds may be challenging, as well.

Acknowledgement

The PURLs Surveillance System is supported in part by Grant Number UL1RR024999 from the National Center for Research Resources, a Clinical Translational Science Award to the University of Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Research Resources or the National Institutes of Health.

Click here to view PURL METHODOLOGY