Andrew Dunn, MD, FACP, FHM

Clinical question: Is the diagnosis of cardiovascular disease (CVD) associated with the risk of subsequent hip fracture?

Background: Osteoporosis and CVD are regarded as independent, age-related conditions. However, recent research suggests that the bone and vascular systems share common regulatory mechanisms. Stroke is a known risk factor for hip fractures, and bisphosphonates have been shown to prevent atherosclerosis and reduce total mortality rate.

Study design: Cohort study.

Setting: Swedish National Patient Registry.

Synopsis: The study identified 31,936 Swedish twins born from 1914 to 1944. This cohort was followed up to age 50, and time-dependent exposures using Cox-proportional hazard regression models were evaluated.

Times to hip fracture after CVD diagnosis were isolated. Crude absolute rate of hip fractures (per 1,000 person-years) was 12.6 after diagnosis of heart failure, 12.6 after a stroke, 6.6 after peripheral atherosclerosis, and 5.2 after ischemic heart disease (IHD), compared with 1.2 per 1,000 person-years without a CVD diagnosis. Multivariable-adjusted hazard ratio (HR) of hip fracture after heart failure was 4.40 (95% CI, 3.43-5.63); after a stroke was 5.09 (95% CI, 4.18-6.20); after peripheral atherosclerosis was 3.20 (CI, 2.28-4.50); and after an IHD event was 2.32 (CI, 1.91-2.84).

Identical twins even without heart failure and stroke also had an increased risk of hip fracture if their twin had been diagnosed with these diseases.

Bottom line: Cardiovascular disease is significantly associated with risk of subsequent hip fracture, and genetic factors probably play a role in the association.

Citation: Sennerby U, Melhus H, Gedeborg R, et al. Cardiovascular diseases and risk of hip fracture. JAMA. 2009;302(15):1666-1673.

Clinical question: Is the diagnosis of cardiovascular disease (CVD) associated with the risk of subsequent hip fracture?

Background: Osteoporosis and CVD are regarded as independent, age-related conditions. However, recent research suggests that the bone and vascular systems share common regulatory mechanisms. Stroke is a known risk factor for hip fractures, and bisphosphonates have been shown to prevent atherosclerosis and reduce total mortality rate.

Study design: Cohort study.

Setting: Swedish National Patient Registry.

Synopsis: The study identified 31,936 Swedish twins born from 1914 to 1944. This cohort was followed up to age 50, and time-dependent exposures using Cox-proportional hazard regression models were evaluated.

Times to hip fracture after CVD diagnosis were isolated. Crude absolute rate of hip fractures (per 1,000 person-years) was 12.6 after diagnosis of heart failure, 12.6 after a stroke, 6.6 after peripheral atherosclerosis, and 5.2 after ischemic heart disease (IHD), compared with 1.2 per 1,000 person-years without a CVD diagnosis. Multivariable-adjusted hazard ratio (HR) of hip fracture after heart failure was 4.40 (95% CI, 3.43-5.63); after a stroke was 5.09 (95% CI, 4.18-6.20); after peripheral atherosclerosis was 3.20 (CI, 2.28-4.50); and after an IHD event was 2.32 (CI, 1.91-2.84).

Identical twins even without heart failure and stroke also had an increased risk of hip fracture if their twin had been diagnosed with these diseases.

Bottom line: Cardiovascular disease is significantly associated with risk of subsequent hip fracture, and genetic factors probably play a role in the association.

Citation: Sennerby U, Melhus H, Gedeborg R, et al. Cardiovascular diseases and risk of hip fracture. JAMA. 2009;302(15):1666-1673.

Clinical question: Is the diagnosis of cardiovascular disease (CVD) associated with the risk of subsequent hip fracture?

Background: Osteoporosis and CVD are regarded as independent, age-related conditions. However, recent research suggests that the bone and vascular systems share common regulatory mechanisms. Stroke is a known risk factor for hip fractures, and bisphosphonates have been shown to prevent atherosclerosis and reduce total mortality rate.

Study design: Cohort study.

Setting: Swedish National Patient Registry.

Synopsis: The study identified 31,936 Swedish twins born from 1914 to 1944. This cohort was followed up to age 50, and time-dependent exposures using Cox-proportional hazard regression models were evaluated.

Times to hip fracture after CVD diagnosis were isolated. Crude absolute rate of hip fractures (per 1,000 person-years) was 12.6 after diagnosis of heart failure, 12.6 after a stroke, 6.6 after peripheral atherosclerosis, and 5.2 after ischemic heart disease (IHD), compared with 1.2 per 1,000 person-years without a CVD diagnosis. Multivariable-adjusted hazard ratio (HR) of hip fracture after heart failure was 4.40 (95% CI, 3.43-5.63); after a stroke was 5.09 (95% CI, 4.18-6.20); after peripheral atherosclerosis was 3.20 (CI, 2.28-4.50); and after an IHD event was 2.32 (CI, 1.91-2.84).

Identical twins even without heart failure and stroke also had an increased risk of hip fracture if their twin had been diagnosed with these diseases.

Bottom line: Cardiovascular disease is significantly associated with risk of subsequent hip fracture, and genetic factors probably play a role in the association.

Citation: Sennerby U, Melhus H, Gedeborg R, et al. Cardiovascular diseases and risk of hip fracture. JAMA. 2009;302(15):1666-1673.

Clinical question: Do over-the-counter (OTC) analgesics lead to acute hepatic decompensation among patients with cirrhosis?

Background: In theory, intake of acetaminophen and/or nonsteroidal anti-inflammatory drugs (NSAIDs) can worsen hepatic function and lead to complications among cirrhotic patients. The role of OTC analgesics in potentially triggering acute hepatic decompensation among cirrhotic patients has not been studied.

Study design: Prospective case-control study.

Setting: Two tertiary-care hospitals.

Synopsis: Cirrhotic patients hospitalized for acute liver decompensation were compared with compensated cirrhotic patients in the liver clinic (cirrhotic controls) and with randomly selected, noncirrhotic patients who were simultaneously hospitalized (noncirrhotic controls). Data collected through questionnaires included quantity and dose of OTC analgesics used and alcohol consumption in the past 30 days.

Thirty-five percent of the hospitalized cirrhotic patients, 52% of the cirrhotic controls, and 70% of the noncirrhotic controls used OTC analgesics. At doses lower than those recommended, acetaminophen is not associated with acute liver decompensation among cirrhotic patients, even with recent alcohol use. However, NSAIDs taken by the cirrhotic patients, when compared to control subjects, were in larger doses and used for a longer duration, suggesting NSAIDs may have contributed to the acute decompensation.

Study limitations include the nature of the study design, reliance on the patient’s recall of OTC analgesic use, and obtaining other possible causes of decompensation, such as herbal supplement intake or compliance with diuretics or dietary indiscretion.

Bottom line: Acetaminophen at doses lower than recommended is not associated with adverse complications in cirrhotic patients, but NSAIDs are possibly associated with acute decompensation.

Citation: Khalid SK, Lane J, Navarro V, Garcia-Tsao G. Use of over-the-counter analgesics is not associated with acute decompensation in patients with cirrhosis. Clin Gastroenterol Hepatol. 2009;7(9):994-999.

Clinical question: Do over-the-counter (OTC) analgesics lead to acute hepatic decompensation among patients with cirrhosis?

Background: In theory, intake of acetaminophen and/or nonsteroidal anti-inflammatory drugs (NSAIDs) can worsen hepatic function and lead to complications among cirrhotic patients. The role of OTC analgesics in potentially triggering acute hepatic decompensation among cirrhotic patients has not been studied.

Study design: Prospective case-control study.

Setting: Two tertiary-care hospitals.

Synopsis: Cirrhotic patients hospitalized for acute liver decompensation were compared with compensated cirrhotic patients in the liver clinic (cirrhotic controls) and with randomly selected, noncirrhotic patients who were simultaneously hospitalized (noncirrhotic controls). Data collected through questionnaires included quantity and dose of OTC analgesics used and alcohol consumption in the past 30 days.

Thirty-five percent of the hospitalized cirrhotic patients, 52% of the cirrhotic controls, and 70% of the noncirrhotic controls used OTC analgesics. At doses lower than those recommended, acetaminophen is not associated with acute liver decompensation among cirrhotic patients, even with recent alcohol use. However, NSAIDs taken by the cirrhotic patients, when compared to control subjects, were in larger doses and used for a longer duration, suggesting NSAIDs may have contributed to the acute decompensation.

Study limitations include the nature of the study design, reliance on the patient’s recall of OTC analgesic use, and obtaining other possible causes of decompensation, such as herbal supplement intake or compliance with diuretics or dietary indiscretion.

Bottom line: Acetaminophen at doses lower than recommended is not associated with adverse complications in cirrhotic patients, but NSAIDs are possibly associated with acute decompensation.

Citation: Khalid SK, Lane J, Navarro V, Garcia-Tsao G. Use of over-the-counter analgesics is not associated with acute decompensation in patients with cirrhosis. Clin Gastroenterol Hepatol. 2009;7(9):994-999.

Clinical question: Do over-the-counter (OTC) analgesics lead to acute hepatic decompensation among patients with cirrhosis?

Background: In theory, intake of acetaminophen and/or nonsteroidal anti-inflammatory drugs (NSAIDs) can worsen hepatic function and lead to complications among cirrhotic patients. The role of OTC analgesics in potentially triggering acute hepatic decompensation among cirrhotic patients has not been studied.

Study design: Prospective case-control study.

Setting: Two tertiary-care hospitals.

Synopsis: Cirrhotic patients hospitalized for acute liver decompensation were compared with compensated cirrhotic patients in the liver clinic (cirrhotic controls) and with randomly selected, noncirrhotic patients who were simultaneously hospitalized (noncirrhotic controls). Data collected through questionnaires included quantity and dose of OTC analgesics used and alcohol consumption in the past 30 days.

Thirty-five percent of the hospitalized cirrhotic patients, 52% of the cirrhotic controls, and 70% of the noncirrhotic controls used OTC analgesics. At doses lower than those recommended, acetaminophen is not associated with acute liver decompensation among cirrhotic patients, even with recent alcohol use. However, NSAIDs taken by the cirrhotic patients, when compared to control subjects, were in larger doses and used for a longer duration, suggesting NSAIDs may have contributed to the acute decompensation.

Study limitations include the nature of the study design, reliance on the patient’s recall of OTC analgesic use, and obtaining other possible causes of decompensation, such as herbal supplement intake or compliance with diuretics or dietary indiscretion.

Bottom line: Acetaminophen at doses lower than recommended is not associated with adverse complications in cirrhotic patients, but NSAIDs are possibly associated with acute decompensation.

Citation: Khalid SK, Lane J, Navarro V, Garcia-Tsao G. Use of over-the-counter analgesics is not associated with acute decompensation in patients with cirrhosis. Clin Gastroenterol Hepatol. 2009;7(9):994-999.

Clinical question: How do perioperative and long-term morbidity and mortality compare in endovascular and open repair of abdominal aortic aneurysm (AAA)?

Background: Open AAA repair has relatively high perioperative mortality. Endovascular repair was developed as a less-invasive option and has been shown to reduce inpatient perioperative mortality, length of hospital stay, and ICU requirement. However, data suggest it leads to more frequent reinterventions and the same mortality rate as open repair at two years.

Study design: Randomized clinical trial.

Setting: Veterans Affairs medical centers.

Synopsis: The study randomized 881 veterans who planned to have elective AAA repair and were eligible for both endovascular and open repair. This is a planned, two-year interim report in a nine-year study.

Perioperative mortality was 0.5% in the endovascular repair group, compared with 3.0% in the open repair group. However, this difference in mortality was not statistically significant at two years. The endovascular repair group experienced shorter procedure and mechanical ventilation time, decreased hospital and ICU stay, and lower rate of blood transfusions.

Overall, there was no difference between the groups for major morbidity, procedure failure, need for secondary therapeutic intervention, quality of life, or erectile dysfunction. More data on long-term comparison of these two interventions will be available at the conclusion of this study.

Bottom line: Endovascular repair of AAA has lower perioperative mortality than open repair but did not lead to improved morbidity or mortality at two years.

Citation: Lederle FA, Freischlag JA, Kyriakides TC, et al. Outcomes following endovascular vs. open repair of abdominal aortic aneurysm: a randomized trial. JAMA. 2009;302 (14):1535-1542.

Clinical question: How do perioperative and long-term morbidity and mortality compare in endovascular and open repair of abdominal aortic aneurysm (AAA)?

Background: Open AAA repair has relatively high perioperative mortality. Endovascular repair was developed as a less-invasive option and has been shown to reduce inpatient perioperative mortality, length of hospital stay, and ICU requirement. However, data suggest it leads to more frequent reinterventions and the same mortality rate as open repair at two years.

Study design: Randomized clinical trial.

Setting: Veterans Affairs medical centers.

Synopsis: The study randomized 881 veterans who planned to have elective AAA repair and were eligible for both endovascular and open repair. This is a planned, two-year interim report in a nine-year study.

Perioperative mortality was 0.5% in the endovascular repair group, compared with 3.0% in the open repair group. However, this difference in mortality was not statistically significant at two years. The endovascular repair group experienced shorter procedure and mechanical ventilation time, decreased hospital and ICU stay, and lower rate of blood transfusions.

Overall, there was no difference between the groups for major morbidity, procedure failure, need for secondary therapeutic intervention, quality of life, or erectile dysfunction. More data on long-term comparison of these two interventions will be available at the conclusion of this study.

Bottom line: Endovascular repair of AAA has lower perioperative mortality than open repair but did not lead to improved morbidity or mortality at two years.

Citation: Lederle FA, Freischlag JA, Kyriakides TC, et al. Outcomes following endovascular vs. open repair of abdominal aortic aneurysm: a randomized trial. JAMA. 2009;302 (14):1535-1542.

Clinical question: How do perioperative and long-term morbidity and mortality compare in endovascular and open repair of abdominal aortic aneurysm (AAA)?

Background: Open AAA repair has relatively high perioperative mortality. Endovascular repair was developed as a less-invasive option and has been shown to reduce inpatient perioperative mortality, length of hospital stay, and ICU requirement. However, data suggest it leads to more frequent reinterventions and the same mortality rate as open repair at two years.

Study design: Randomized clinical trial.

Setting: Veterans Affairs medical centers.

Synopsis: The study randomized 881 veterans who planned to have elective AAA repair and were eligible for both endovascular and open repair. This is a planned, two-year interim report in a nine-year study.

Perioperative mortality was 0.5% in the endovascular repair group, compared with 3.0% in the open repair group. However, this difference in mortality was not statistically significant at two years. The endovascular repair group experienced shorter procedure and mechanical ventilation time, decreased hospital and ICU stay, and lower rate of blood transfusions.

Overall, there was no difference between the groups for major morbidity, procedure failure, need for secondary therapeutic intervention, quality of life, or erectile dysfunction. More data on long-term comparison of these two interventions will be available at the conclusion of this study.

Bottom line: Endovascular repair of AAA has lower perioperative mortality than open repair but did not lead to improved morbidity or mortality at two years.

Citation: Lederle FA, Freischlag JA, Kyriakides TC, et al. Outcomes following endovascular vs. open repair of abdominal aortic aneurysm: a randomized trial. JAMA. 2009;302 (14):1535-1542.

Clinical question: Does perioperative beta-blocker discontinuation affect postoperative myocardial infarction (MI) in low-risk patients undergoing joint arthroplasty?

Background: Recent trials show no benefit of perioperative beta blockers in reducing the incidence of perioperative myocardial infarctions (POMI) in low-risk patients. This retrospective study examined the impact of continuing or discontinuing beta blockers and the occurrence of POMI in patients undergoing elective joint arthroplasties.

Study design: Retrospective chart review.

Setting: Large academic center in Ottawa, Canada.

Synopsis: Medical records for 5,178 patients undergoing elective hip or knee arthroplasty from January 2002 to June 2006 were included in the review. The primary outcome was POMI, defined as an increased troponin level. Patients were divided into three groups: beta blocker prescribed on post-operative day (POD) zero and continued for one week or until discharge; beta blocker prescribed on POD zero and discontinued at any time in the first week; and no beta blocker on POD 0.

Beta blockers were continued in 992 patients and discontinued in 252 patients. The rate of POMI and death increased in the beta-blocker discontinuation group (odds ratio 2.0 [1.1-3.9] and 2.0 [1.1-3.9], respectively). This association persisted after adjustment for cardiac risk using a validated risk score.

The study was limited by the fact that the control group did not include patients who were on a beta blocker at home, thus potentially increasing the number of events in this group. The discontinuation beta blocker group had an increased baseline risk for POMI. The reason for discontinuing the beta blocker was not known, and cessation of beta blocker could have been due to an acute event.

Bottom line: This study adds support to the American College of Cardiology and American Heart Association (ACC/AHA) guidelines, which recommend continuation of beta-blocker therapy in the perioperative period.

Citation: Van Klei WA, Bryson GL, Yang H, Forster AJ. Effect of beta-blocker prescription on the incidence of postoperative myocardial infarction after hip and knee arthroplasty. Anesthesiology. 2009;111(4):717-724.

Clinical question: Does perioperative beta-blocker discontinuation affect postoperative myocardial infarction (MI) in low-risk patients undergoing joint arthroplasty?

Background: Recent trials show no benefit of perioperative beta blockers in reducing the incidence of perioperative myocardial infarctions (POMI) in low-risk patients. This retrospective study examined the impact of continuing or discontinuing beta blockers and the occurrence of POMI in patients undergoing elective joint arthroplasties.

Study design: Retrospective chart review.

Setting: Large academic center in Ottawa, Canada.

Synopsis: Medical records for 5,178 patients undergoing elective hip or knee arthroplasty from January 2002 to June 2006 were included in the review. The primary outcome was POMI, defined as an increased troponin level. Patients were divided into three groups: beta blocker prescribed on post-operative day (POD) zero and continued for one week or until discharge; beta blocker prescribed on POD zero and discontinued at any time in the first week; and no beta blocker on POD 0.

Beta blockers were continued in 992 patients and discontinued in 252 patients. The rate of POMI and death increased in the beta-blocker discontinuation group (odds ratio 2.0 [1.1-3.9] and 2.0 [1.1-3.9], respectively). This association persisted after adjustment for cardiac risk using a validated risk score.

The study was limited by the fact that the control group did not include patients who were on a beta blocker at home, thus potentially increasing the number of events in this group. The discontinuation beta blocker group had an increased baseline risk for POMI. The reason for discontinuing the beta blocker was not known, and cessation of beta blocker could have been due to an acute event.

Bottom line: This study adds support to the American College of Cardiology and American Heart Association (ACC/AHA) guidelines, which recommend continuation of beta-blocker therapy in the perioperative period.

Citation: Van Klei WA, Bryson GL, Yang H, Forster AJ. Effect of beta-blocker prescription on the incidence of postoperative myocardial infarction after hip and knee arthroplasty. Anesthesiology. 2009;111(4):717-724.

Clinical question: Does perioperative beta-blocker discontinuation affect postoperative myocardial infarction (MI) in low-risk patients undergoing joint arthroplasty?

Background: Recent trials show no benefit of perioperative beta blockers in reducing the incidence of perioperative myocardial infarctions (POMI) in low-risk patients. This retrospective study examined the impact of continuing or discontinuing beta blockers and the occurrence of POMI in patients undergoing elective joint arthroplasties.

Study design: Retrospective chart review.

Setting: Large academic center in Ottawa, Canada.

Synopsis: Medical records for 5,178 patients undergoing elective hip or knee arthroplasty from January 2002 to June 2006 were included in the review. The primary outcome was POMI, defined as an increased troponin level. Patients were divided into three groups: beta blocker prescribed on post-operative day (POD) zero and continued for one week or until discharge; beta blocker prescribed on POD zero and discontinued at any time in the first week; and no beta blocker on POD 0.

Beta blockers were continued in 992 patients and discontinued in 252 patients. The rate of POMI and death increased in the beta-blocker discontinuation group (odds ratio 2.0 [1.1-3.9] and 2.0 [1.1-3.9], respectively). This association persisted after adjustment for cardiac risk using a validated risk score.

The study was limited by the fact that the control group did not include patients who were on a beta blocker at home, thus potentially increasing the number of events in this group. The discontinuation beta blocker group had an increased baseline risk for POMI. The reason for discontinuing the beta blocker was not known, and cessation of beta blocker could have been due to an acute event.

Bottom line: This study adds support to the American College of Cardiology and American Heart Association (ACC/AHA) guidelines, which recommend continuation of beta-blocker therapy in the perioperative period.

Citation: Van Klei WA, Bryson GL, Yang H, Forster AJ. Effect of beta-blocker prescription on the incidence of postoperative myocardial infarction after hip and knee arthroplasty. Anesthesiology. 2009;111(4):717-724.

Clinical question: On what aspects of VTE management in cancer patients are there consensus among the major guideline panels?

Background: VTE is a common and serious complication of cancer. Patients might be hypercoagulable due to prothrombotic mediators released or mediated by tumor cells, chemotherapeutic agents, debility, central venous catheters, hospitalizations, or surgical procedures. The optimal management often is problematic due to uncertain benefit and risk of bleeding.

Study design: Review of major guideline statements.

Synopsis: The authors examined five VTE guidelines of American and European cancer societies. Each guideline was reviewed to determine the main recommendations and whether there was consensus on key aspects of anticoagulant management.

The study authors concluded that consensus was reached on most key recommendations:

• VTE prophylaxis in hospitalized medical patients. All five guidelines recommend the use of prophylaxis, though some guidelines recommend anticoagulant prophylaxis for all inpatients in the absence of contraindications and some recommend limiting prophylaxis to immobilized patients. All five recommend the use of either unfractionated heparin, low-molecular-weight heparin (LMWH), or fondaparinux.
• VTE prevention in cancer patients undergoing surgery. All five guidelines recommend anticoagulant prophylaxis in the absence of contraindications and extending prophylaxis approximately four weeks after major surgery.
• VTE prophylaxis in cancer patients with central venous catheters. Not recommended.
• VTE prophylaxis in ambulatory cancer patients without central venous catheters. Recommended only for multiple myeloma patients receiving a thalidomide-lenalidomide regimen.
• Long-term treatment of acute VTE in cancer patients. All five guidelines recommend initial treatment with LMWH for at least three to six months, followed by indefinite treatment with LMWH or a vitamin K antagonist.

Bottom line: Major guideline panels agree on key aspects of VTE management for cancer patients, including the use of prophylaxis for hospitalized medical and surgical patients and the use of long-term LMWH treatment for cancer patients with acute VTE.

Citation: Khorana AA, Streiff MB, Farge D, et al. Venous thromboembolism prophylaxis and treatment in cancer: a consensus statement of major guidelines panels and call to action. J Clin Oncol. 2009; 27(29):4919-4926.

Clinical question: On what aspects of VTE management in cancer patients are there consensus among the major guideline panels?

Background: VTE is a common and serious complication of cancer. Patients might be hypercoagulable due to prothrombotic mediators released or mediated by tumor cells, chemotherapeutic agents, debility, central venous catheters, hospitalizations, or surgical procedures. The optimal management often is problematic due to uncertain benefit and risk of bleeding.

Study design: Review of major guideline statements.

Synopsis: The authors examined five VTE guidelines of American and European cancer societies. Each guideline was reviewed to determine the main recommendations and whether there was consensus on key aspects of anticoagulant management.

The study authors concluded that consensus was reached on most key recommendations:

• VTE prophylaxis in hospitalized medical patients. All five guidelines recommend the use of prophylaxis, though some guidelines recommend anticoagulant prophylaxis for all inpatients in the absence of contraindications and some recommend limiting prophylaxis to immobilized patients. All five recommend the use of either unfractionated heparin, low-molecular-weight heparin (LMWH), or fondaparinux.
• VTE prevention in cancer patients undergoing surgery. All five guidelines recommend anticoagulant prophylaxis in the absence of contraindications and extending prophylaxis approximately four weeks after major surgery.
• VTE prophylaxis in cancer patients with central venous catheters. Not recommended.
• VTE prophylaxis in ambulatory cancer patients without central venous catheters. Recommended only for multiple myeloma patients receiving a thalidomide-lenalidomide regimen.
• Long-term treatment of acute VTE in cancer patients. All five guidelines recommend initial treatment with LMWH for at least three to six months, followed by indefinite treatment with LMWH or a vitamin K antagonist.

Bottom line: Major guideline panels agree on key aspects of VTE management for cancer patients, including the use of prophylaxis for hospitalized medical and surgical patients and the use of long-term LMWH treatment for cancer patients with acute VTE.

Citation: Khorana AA, Streiff MB, Farge D, et al. Venous thromboembolism prophylaxis and treatment in cancer: a consensus statement of major guidelines panels and call to action. J Clin Oncol. 2009; 27(29):4919-4926.

Clinical question: On what aspects of VTE management in cancer patients are there consensus among the major guideline panels?

Background: VTE is a common and serious complication of cancer. Patients might be hypercoagulable due to prothrombotic mediators released or mediated by tumor cells, chemotherapeutic agents, debility, central venous catheters, hospitalizations, or surgical procedures. The optimal management often is problematic due to uncertain benefit and risk of bleeding.

Study design: Review of major guideline statements.

Synopsis: The authors examined five VTE guidelines of American and European cancer societies. Each guideline was reviewed to determine the main recommendations and whether there was consensus on key aspects of anticoagulant management.

The study authors concluded that consensus was reached on most key recommendations:

• VTE prophylaxis in hospitalized medical patients. All five guidelines recommend the use of prophylaxis, though some guidelines recommend anticoagulant prophylaxis for all inpatients in the absence of contraindications and some recommend limiting prophylaxis to immobilized patients. All five recommend the use of either unfractionated heparin, low-molecular-weight heparin (LMWH), or fondaparinux.
• VTE prevention in cancer patients undergoing surgery. All five guidelines recommend anticoagulant prophylaxis in the absence of contraindications and extending prophylaxis approximately four weeks after major surgery.
• VTE prophylaxis in cancer patients with central venous catheters. Not recommended.
• VTE prophylaxis in ambulatory cancer patients without central venous catheters. Recommended only for multiple myeloma patients receiving a thalidomide-lenalidomide regimen.
• Long-term treatment of acute VTE in cancer patients. All five guidelines recommend initial treatment with LMWH for at least three to six months, followed by indefinite treatment with LMWH or a vitamin K antagonist.

Bottom line: Major guideline panels agree on key aspects of VTE management for cancer patients, including the use of prophylaxis for hospitalized medical and surgical patients and the use of long-term LMWH treatment for cancer patients with acute VTE.

Citation: Khorana AA, Streiff MB, Farge D, et al. Venous thromboembolism prophylaxis and treatment in cancer: a consensus statement of major guidelines panels and call to action. J Clin Oncol. 2009; 27(29):4919-4926.

Clinical question: What is the relationship between acute DVT and pulmonary embolism (PE) in trauma patients?

Background: Major trauma is associated with an increased risk of acute DVT and PE. It is assumed that the majority of PEs arise from DVTs in the lower extremities. Definitive evidence demonstrating that PEs form in situ rather than embolize from leg veins could impact indications for inferior vena cava filters.

Study design: Retrospective chart review.

Setting: Academic Level 1 trauma center in Boston.

Synopsis: The medical records of 247 trauma patients with suspected PE who underwent CT angiography of the lungs and simultaneous CT venography of the pelvis and lower extremities from January 2004 to December 2007 were reviewed. High-risk patients also underwent weekly screening with duplex ultrasonagraphy of the legs.

PE was diagnosed in 46 patients (19%) and DVT in 18 patients (7%). Anticoagulant prophylaxis had been administered to 96% and 78% of the patients with PE and DVT, respectively. PE was diagnosed a median of 5.5 days after admission (range 0-40 days) and the majority (61%) were in segmental or subsegmental branches, rather than in the main or lobar pulmonary arteries (39%). Only seven of the 46 patients (15%) diagnosed with PE also had a pelvic or lower-extremity DVT on simultaneous imaging with CT venography.

Bottom line: Trauma patients with PE often do not have a DVT at the time of diagnosis, though it remains unknown whether this is due to in-situ pulmonary thrombosis or complete embolization from the lower extremities.

Citation: Velmahos GC, Spaniolas K, Tabbara M, et al. Pulmonary embolism and deep venous thrombosis in trauma: are they related? Arch Surg. 2009;144:928-932.

Clinical question: What is the relationship between acute DVT and pulmonary embolism (PE) in trauma patients?

Background: Major trauma is associated with an increased risk of acute DVT and PE. It is assumed that the majority of PEs arise from DVTs in the lower extremities. Definitive evidence demonstrating that PEs form in situ rather than embolize from leg veins could impact indications for inferior vena cava filters.

Study design: Retrospective chart review.

Setting: Academic Level 1 trauma center in Boston.

Synopsis: The medical records of 247 trauma patients with suspected PE who underwent CT angiography of the lungs and simultaneous CT venography of the pelvis and lower extremities from January 2004 to December 2007 were reviewed. High-risk patients also underwent weekly screening with duplex ultrasonagraphy of the legs.

PE was diagnosed in 46 patients (19%) and DVT in 18 patients (7%). Anticoagulant prophylaxis had been administered to 96% and 78% of the patients with PE and DVT, respectively. PE was diagnosed a median of 5.5 days after admission (range 0-40 days) and the majority (61%) were in segmental or subsegmental branches, rather than in the main or lobar pulmonary arteries (39%). Only seven of the 46 patients (15%) diagnosed with PE also had a pelvic or lower-extremity DVT on simultaneous imaging with CT venography.

Bottom line: Trauma patients with PE often do not have a DVT at the time of diagnosis, though it remains unknown whether this is due to in-situ pulmonary thrombosis or complete embolization from the lower extremities.

Citation: Velmahos GC, Spaniolas K, Tabbara M, et al. Pulmonary embolism and deep venous thrombosis in trauma: are they related? Arch Surg. 2009;144:928-932.

Clinical question: What is the relationship between acute DVT and pulmonary embolism (PE) in trauma patients?

Background: Major trauma is associated with an increased risk of acute DVT and PE. It is assumed that the majority of PEs arise from DVTs in the lower extremities. Definitive evidence demonstrating that PEs form in situ rather than embolize from leg veins could impact indications for inferior vena cava filters.

Study design: Retrospective chart review.

Setting: Academic Level 1 trauma center in Boston.

Synopsis: The medical records of 247 trauma patients with suspected PE who underwent CT angiography of the lungs and simultaneous CT venography of the pelvis and lower extremities from January 2004 to December 2007 were reviewed. High-risk patients also underwent weekly screening with duplex ultrasonagraphy of the legs.

PE was diagnosed in 46 patients (19%) and DVT in 18 patients (7%). Anticoagulant prophylaxis had been administered to 96% and 78% of the patients with PE and DVT, respectively. PE was diagnosed a median of 5.5 days after admission (range 0-40 days) and the majority (61%) were in segmental or subsegmental branches, rather than in the main or lobar pulmonary arteries (39%). Only seven of the 46 patients (15%) diagnosed with PE also had a pelvic or lower-extremity DVT on simultaneous imaging with CT venography.

Bottom line: Trauma patients with PE often do not have a DVT at the time of diagnosis, though it remains unknown whether this is due to in-situ pulmonary thrombosis or complete embolization from the lower extremities.

Citation: Velmahos GC, Spaniolas K, Tabbara M, et al. Pulmonary embolism and deep venous thrombosis in trauma: are they related? Arch Surg. 2009;144:928-932.

Clinical question: Using available data, what is the estimated cost savings of eliminating adverse events and avoiding redundant tests?

Background: Reimbursement schemes are changing such that hospitals are reimbursed less for some adverse events. This financial disincentive is expected to spark interest in improved patient safety. The authors sought to model the cost savings generated by eliminating redundant testing and adverse events from literature-based estimates.

Study design: Development of conceptual model to identify common or costly adverse events, redundant tests, and simulated costs.

Setting: Literature review, expert opinion, data from safety organizations and epidemiologic studies, and patient data from the 2004 National Inpatient Data Sample.

Synopsis: The conceptual model identified 5.7 million adverse events in U.S. hospitals, of which 3 million were considered preventable. The most common events included hospital-acquired infections (82% preventable), adverse drug events (26%), falls (33%), and iatrogenic thromboembolic events (62%). The calculated cost savings totaled $16.6 billion (5.5% of total inpatient costs) for adverse events and $8.2 billion for the elimination of redundant tests. When looking at hospital subtypes, the greatest savings would come from major teaching hospitals.

This study is limited by its use of published and heterogeneous data spanning a 15-year period. The authors did not include events for which there was no epidemiologic or cost data. As hospital-care changes and technology is adopted, it is uncertain how this changes the costs, prevalence, and the preventable nature of these events. The model was not consistently able to identifying high- and low-risk patients. For instance, in some models, all patients were considered at risk for events.

Bottom line: Based on a conceptual model of 2004 hospitalized patients, eliminating preventable adverse events could have saved $16.6 billion, while eliminating redundant tests could have saved another $8 billion.

Citation: Jha AK, Chan DC, Ridgway AB, Franz C, Bates DW. Improving safety and eliminating redundant tests: cutting costs in U.S. hospitals. Health Aff (Millwood). 2009;28(5):1475-1484.

Clinical question: Using available data, what is the estimated cost savings of eliminating adverse events and avoiding redundant tests?

Background: Reimbursement schemes are changing such that hospitals are reimbursed less for some adverse events. This financial disincentive is expected to spark interest in improved patient safety. The authors sought to model the cost savings generated by eliminating redundant testing and adverse events from literature-based estimates.

Study design: Development of conceptual model to identify common or costly adverse events, redundant tests, and simulated costs.

Setting: Literature review, expert opinion, data from safety organizations and epidemiologic studies, and patient data from the 2004 National Inpatient Data Sample.

Synopsis: The conceptual model identified 5.7 million adverse events in U.S. hospitals, of which 3 million were considered preventable. The most common events included hospital-acquired infections (82% preventable), adverse drug events (26%), falls (33%), and iatrogenic thromboembolic events (62%). The calculated cost savings totaled $16.6 billion (5.5% of total inpatient costs) for adverse events and $8.2 billion for the elimination of redundant tests. When looking at hospital subtypes, the greatest savings would come from major teaching hospitals.

This study is limited by its use of published and heterogeneous data spanning a 15-year period. The authors did not include events for which there was no epidemiologic or cost data. As hospital-care changes and technology is adopted, it is uncertain how this changes the costs, prevalence, and the preventable nature of these events. The model was not consistently able to identifying high- and low-risk patients. For instance, in some models, all patients were considered at risk for events.

Bottom line: Based on a conceptual model of 2004 hospitalized patients, eliminating preventable adverse events could have saved $16.6 billion, while eliminating redundant tests could have saved another $8 billion.

Citation: Jha AK, Chan DC, Ridgway AB, Franz C, Bates DW. Improving safety and eliminating redundant tests: cutting costs in U.S. hospitals. Health Aff (Millwood). 2009;28(5):1475-1484.

Clinical question: Using available data, what is the estimated cost savings of eliminating adverse events and avoiding redundant tests?

Background: Reimbursement schemes are changing such that hospitals are reimbursed less for some adverse events. This financial disincentive is expected to spark interest in improved patient safety. The authors sought to model the cost savings generated by eliminating redundant testing and adverse events from literature-based estimates.

Study design: Development of conceptual model to identify common or costly adverse events, redundant tests, and simulated costs.

Setting: Literature review, expert opinion, data from safety organizations and epidemiologic studies, and patient data from the 2004 National Inpatient Data Sample.

Synopsis: The conceptual model identified 5.7 million adverse events in U.S. hospitals, of which 3 million were considered preventable. The most common events included hospital-acquired infections (82% preventable), adverse drug events (26%), falls (33%), and iatrogenic thromboembolic events (62%). The calculated cost savings totaled $16.6 billion (5.5% of total inpatient costs) for adverse events and $8.2 billion for the elimination of redundant tests. When looking at hospital subtypes, the greatest savings would come from major teaching hospitals.

This study is limited by its use of published and heterogeneous data spanning a 15-year period. The authors did not include events for which there was no epidemiologic or cost data. As hospital-care changes and technology is adopted, it is uncertain how this changes the costs, prevalence, and the preventable nature of these events. The model was not consistently able to identifying high- and low-risk patients. For instance, in some models, all patients were considered at risk for events.

Bottom line: Based on a conceptual model of 2004 hospitalized patients, eliminating preventable adverse events could have saved $16.6 billion, while eliminating redundant tests could have saved another $8 billion.

Citation: Jha AK, Chan DC, Ridgway AB, Franz C, Bates DW. Improving safety and eliminating redundant tests: cutting costs in U.S. hospitals. Health Aff (Millwood). 2009;28(5):1475-1484.

Clinical question: Does the use of 80% oxygen perioperatively in abdominal surgery decrease the frequency of surgical-site infection within 14 days without increasing the rate of pulmonary complications?

Background: Low oxygen tension in wounds can negatively impact immune response and healing. Increasing inspiratory oxygen fraction during the perioperative period translates into higher wound oxygen tension. However, the benefit of increased oxygen fraction therapy in abdominal surgery healing and complications is not clear, nor is the frequency of pulmonary complications.

Study design: Patient- and observer-blinded clinical trial.

Setting: Fourteen Danish hospitals from October 2006 to October 2008.

Synopsis: Patients were randomized to receive a fraction of inspired oxygen (FIO₂) of 0.80 or 0.30. The primary outcome—surgical-site infection in the superficial or deep wound or intra-abdominal cavity within 14 days of surgery—was defined using Centers for Disease Control and Prevention (CDC) criteria. Secondary outcomes included pulmonary complications within 14 days (pneumonia, atelectasis, or respiratory failure), 30-day mortality, duration of post-op course, ICU stay within 14 days post-op, and any abdominal operation within 14 days. The 1,386 patients were enrolled in the intention-to-treat analysis.

Infection occurred in 19.1% of patients given 0.80 FIO2 and in 20.1% of patients given 0.30 FIO2; odds ratio of 0.94 (95% CI 0.72 to 1.22; P=0.64). Numbers of pulmonary complications were not significantly different between the groups.

This trial included acute and nonacute laparotomies with followup for adverse outcomes. Study limitations included the inability to ensure that both groups received timely antibiotics and prevention for hypothermia. Of patients in the 30% FIO2 group, 7.3% required higher oxygen administration. Additionally, infection might have been underestimated in 11.3% of patients who were not followed up on between days 13 and 30.

Bottom line: High oxygen concentration administered during and after laparotomy did not lead to fewer surgical site infections, nor did it significantly increase the frequency of pulmonary complications or death.

Citation: Meyhoff CS, Wetterslev J, Jorgensen LN, et al. Effect of high perioperative oxygen fraction on surgical site infection and pulmonary complications after abdominal surgery: the PROXI randomized clinical trial. JAMA. 2009;302(14):1543-1550.

Clinical question: Does the use of 80% oxygen perioperatively in abdominal surgery decrease the frequency of surgical-site infection within 14 days without increasing the rate of pulmonary complications?

Background: Low oxygen tension in wounds can negatively impact immune response and healing. Increasing inspiratory oxygen fraction during the perioperative period translates into higher wound oxygen tension. However, the benefit of increased oxygen fraction therapy in abdominal surgery healing and complications is not clear, nor is the frequency of pulmonary complications.

Study design: Patient- and observer-blinded clinical trial.

Setting: Fourteen Danish hospitals from October 2006 to October 2008.

Synopsis: Patients were randomized to receive a fraction of inspired oxygen (FIO₂) of 0.80 or 0.30. The primary outcome—surgical-site infection in the superficial or deep wound or intra-abdominal cavity within 14 days of surgery—was defined using Centers for Disease Control and Prevention (CDC) criteria. Secondary outcomes included pulmonary complications within 14 days (pneumonia, atelectasis, or respiratory failure), 30-day mortality, duration of post-op course, ICU stay within 14 days post-op, and any abdominal operation within 14 days. The 1,386 patients were enrolled in the intention-to-treat analysis.

Infection occurred in 19.1% of patients given 0.80 FIO2 and in 20.1% of patients given 0.30 FIO2; odds ratio of 0.94 (95% CI 0.72 to 1.22; P=0.64). Numbers of pulmonary complications were not significantly different between the groups.

This trial included acute and nonacute laparotomies with followup for adverse outcomes. Study limitations included the inability to ensure that both groups received timely antibiotics and prevention for hypothermia. Of patients in the 30% FIO2 group, 7.3% required higher oxygen administration. Additionally, infection might have been underestimated in 11.3% of patients who were not followed up on between days 13 and 30.

Bottom line: High oxygen concentration administered during and after laparotomy did not lead to fewer surgical site infections, nor did it significantly increase the frequency of pulmonary complications or death.

Citation: Meyhoff CS, Wetterslev J, Jorgensen LN, et al. Effect of high perioperative oxygen fraction on surgical site infection and pulmonary complications after abdominal surgery: the PROXI randomized clinical trial. JAMA. 2009;302(14):1543-1550.

Clinical question: Does the use of 80% oxygen perioperatively in abdominal surgery decrease the frequency of surgical-site infection within 14 days without increasing the rate of pulmonary complications?

Background: Low oxygen tension in wounds can negatively impact immune response and healing. Increasing inspiratory oxygen fraction during the perioperative period translates into higher wound oxygen tension. However, the benefit of increased oxygen fraction therapy in abdominal surgery healing and complications is not clear, nor is the frequency of pulmonary complications.

Study design: Patient- and observer-blinded clinical trial.

Setting: Fourteen Danish hospitals from October 2006 to October 2008.

Synopsis: Patients were randomized to receive a fraction of inspired oxygen (FIO₂) of 0.80 or 0.30. The primary outcome—surgical-site infection in the superficial or deep wound or intra-abdominal cavity within 14 days of surgery—was defined using Centers for Disease Control and Prevention (CDC) criteria. Secondary outcomes included pulmonary complications within 14 days (pneumonia, atelectasis, or respiratory failure), 30-day mortality, duration of post-op course, ICU stay within 14 days post-op, and any abdominal operation within 14 days. The 1,386 patients were enrolled in the intention-to-treat analysis.

Infection occurred in 19.1% of patients given 0.80 FIO2 and in 20.1% of patients given 0.30 FIO2; odds ratio of 0.94 (95% CI 0.72 to 1.22; P=0.64). Numbers of pulmonary complications were not significantly different between the groups.

This trial included acute and nonacute laparotomies with followup for adverse outcomes. Study limitations included the inability to ensure that both groups received timely antibiotics and prevention for hypothermia. Of patients in the 30% FIO2 group, 7.3% required higher oxygen administration. Additionally, infection might have been underestimated in 11.3% of patients who were not followed up on between days 13 and 30.

Bottom line: High oxygen concentration administered during and after laparotomy did not lead to fewer surgical site infections, nor did it significantly increase the frequency of pulmonary complications or death.

Citation: Meyhoff CS, Wetterslev J, Jorgensen LN, et al. Effect of high perioperative oxygen fraction on surgical site infection and pulmonary complications after abdominal surgery: the PROXI randomized clinical trial. JAMA. 2009;302(14):1543-1550.

In This Edition

Literature at a Glance

A guide to this month’s studies

• Effects of extended VTE prophylaxis in medical patients
• Outcomes with and without preprocedural statins
• Association of subclinical hypothyroidism and CHD
• BP treatment after intracerebral hemorrhage
• Outcomes of ICD therapy in the elderly
• Systems delays and outcomes of STEMI
• D-dimer to predict recurrent VTE
• Stocking height and risk of post-stroke DVT

Extending Anticoagulant Prophylaxis after Medical Hospitalization Decreases VTE, Increases Major Bleeding

Clinical question: For patients with acute medical illness, does extending low-molecular-weight heparin (LMWH) administration for up to 28 days after discharge reduce the incidence of venous thromboembolism (VTE)?

Background: DVT and pulmonary embolism (PE) are common hospital-acquired complications. LMWH has been shown to reduce VTE for medical and surgical patients, and extended-duration LMWH reduces VTE in high-risk surgical patients. Whether extending anticoagulant prophylaxis after discharge for acutely ill medical patients with reduced mobility improves outcomes is unknown.

Study design: Randomized, placebo-controlled trial.

Setting: Three hundred seventy hospitals in 20 countries.

Synopsis: Eligible patients were >40 years old, hospitalized with acute medical illness, and had reduced mobility for ≥3 days. Patients received enoxaparin 40 mg SC daily prophylaxis while hospitalized and were then randomized to an additional 28±4 days of enoxaparin or placebo. Patients received a screening ultrasound to assess for asymptomatic DVT. The primary outcome was a composite of asymptomatic proximal DVT, symptomatic DVT or PE, or fatal PE during the period of extended prophylaxis.

An interim analysis indicated that extended prophylaxis was ineffective; at that time, the protocol was amended to target patients with severe immobility or with moderate immobility plus an additional risk factor (e.g. cancer, prior VTE, or age >75).

The study found that extended prophylaxis decreased the composite VTE outcome (2.5% vs 4.0%, P<0.05) and symptomatic VTE (0.2% vs 1.0%, P<0.05). The incidence of major bleeding was increased in the extended prophylaxis group (0.8% vs 0.3%, P<0.05). There was no difference in mortality.

The unplanned, midstudy protocol amendment to target higher-risk patients is a concern, though the final analyses included patients pre- and post-amendment.

Bottom line: Extending LMWH beyond hospitalization for patients admitted with acute medical illness and decreased mobility decreases VTE, but increases major bleeding.

Citation: Hull RD, Schellong SM, Tapson VF, et al. Extended-duration venous thromboembolism prophylaxis in acutely ill medical patients with recently reduced mobility: a randomized trial. Ann Intern Med. 2010;153(1);8-18.

Preprocedural Statin Therapy Reduces Postprocedural Myocardial Infarction

Clinical question: Does statin therapy reduce periprocedural cardiovascular events?

Background: Myocardial infarction (MI) and death are inherent risks of invasive procedures. Reduction of these risks in certain patient populations has been shown with the use of a beta blockade. Statins have shown promise during acute coronary syndrome. Questions remain about the role of statin therapy before invasive procedures in reducing adverse cardiovascular events.

Study design: Meta-analysis of randomized controlled trials.

Setting: Twenty-one studies involving 4,805 patients, published from inception of MEDLINE, Cochrane, and Clinicaltrials to February 2010.

Synopsis: The use of statins one to seven days preprocedure significantly reduced post-procedural MI in percutaneous coronary interventions (PCI) (P<0.0001). Statins given approximately four weeks in advance of noncardiac surgical procedures also significantly reduced postprocedural MI (P=0.004). An absolute risk reduction of 5.8% for postprocedural MI was found after PCI and 4.1% in noncardiac surgical procedures.

Statins did not show a significant reduction in postprocedural MI (P=0.40) or all-cause mortality (P=0.15) in coronary artery bypass graft surgery (CABG). However, statins did reduce post-CABG atrial fibrillation (P<0.0001).

The 21 studies used a variety of drugs and doses. However, the PCI studies favored atorvastatin 40 mg; more than half the CABG studies used atorvastatin 20 mg; and 91% of the noncardiac surgical studies used fluvastatin 80 mg. Dedicated trials are needed to demonstrate optimal statin agent, dose, and timing of therapy.

Bottom line: Preprocedural statin therapy reduces postprocedural MI after both PCI and noncardiac procedures but not after CABG.

Citation: Winchester DE, Wen X, Xie L, Bavry AA. Evidence of pre-procedural statin therapy: a meta-analysis of randomized trials. J Am Coll Cardiol. 2010;56(19); 1099-1109.

Subclinical Hypothyroidism Increases the Risk of Coronary Heart Disease and Mortality

Clinical question: What are the risks of coronary heart disease (CHD) and mortality among adults with subclinical hypothyroidism?

Background: Subclinical hypo-thyroidism is defined as an elevated serum thyroid stimulating hormone (TSH) level with a normal T4 concentration. Controversy exists regarding the treatment of subclinical hypothyroidism. Because of the association with hyperlipidemia and atherosclerosis, treatment of subclinical hypothyroidism is thought to be beneficial. Previous data from large prospective cohort studies regarding this association are conflicting.

Study design: Study-level meta-analysis of prospective cohort studies.

Setting: Eleven prospective cohorts in the U.S., Europe, Australia, Brazil, and Japan from 1972 to 2007.

Synopsis: Among 55,287 adults, 3,450 (6.2%) had subclinical hypothyroidism and 51,837 were euthyroid. Using Cox proportional hazard models, the association of subclinical hypothyroidism with CHD and mortality were determined for each cohort.

The risk of CHD events and CHD mortality increased with higher TSH concentrations.

In age- and sex-adjusted analyses, the hazard ratio (HR) for CHD events were as follows: HR=1.0 (TSH=4.5-6.9 mIU/L); HR=1.17 (TSH=7-9.9 mIU/L), and HR=1.89 (TSH=10-19.9 mIU/L). Similarly, HRs for CHD mortality showed an increasing trend: 1.09, 1.42, and 1.58, respectively.

Although the association is clearly established here, randomized controlled trials are needed to address whether thyroxine replacement can prevent CHD and the TSH threshold that will provide the most clinical benefit.

Bottom line: Subclinical hypo-thyroidism is associated with an increased risk for CHD events and mortality, primarily in patients with TSH concentrations of 10 mIU/L or higher.

Citation: Rodondi N, den Elzen WP, Bauer DC, et al. Subclinical hypothyroidism and the risk of coronary heart disease and mortality. JAMA. 2010;304(12): 1365-1374.

Reduction in Hematoma Growth after Acute Intracerebral Hemorrhage Associated with Lower Blood Pressure

Clinical question: Does intensive systolic blood pressure (SBP) <140 mmHg within one hour reduce hematoma growth after acute intracerebral bleeding?

Background: Early elevation of blood pressure after an acute intracerebral bleed is strongly associated with hematoma growth and worse outcomes. The pilot phase of the Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial (INTERACT) showed that early intensive blood pressure lowering reduced hematoma growth within six hours after onset of intracerebral hemorrhage.

Study design: Randomized controlled trial in which patients received an early intensive blood-pressure-lowering treatment (goal SBP<140 mmHg within one hour) or the AHA recommended best practice guidelines (goal SBP <180 mmHg), which were published in 1999.

Setting: Network of hospitals in China, South Korea, and Australia.

Synopsis: From 2005 to 2007, the study examined 404 patients with computed-tomography (CT) evidence of intracerebral hemorrhage, elevated SBP (150 mmHg-220 mmHg), and ability to commence BP lowering treatment within six hours of onset. Hematoma volumes were measured in the intensive treatment and guideline groups based on CT scans done at baseline and after 24 hours.

No significant association was found between the baseline SBP and the absolute or proportional growth of the hematoma. However, in the intensively treated patients who achieved target SBP within the first 24 hours, a significant association was found with the absolute and proportional hematoma growth.

Maximum reduction of hematoma growth occurred in the group with a median SBP of 135 mmHg.

This single-observational study did not include patients with severe intracranial bleeding who died or required surgical evacuation within the first 24 hours.

Hematoma size reduction did not improve survival or outcomes.

Bottom line: Intensive SBP lowering from 140 mmHg to 130 mmHg within one hour reduces hematoma growth after an intracranial hemorrhage.

Citation: Arima H, Anderson CS, Wang JG, et al. Lower treatment blood pressure is associated with the greatest reduction in hematoma growth after acute intracerebral hemorrhage. Hypertension. 2010;56(5):852-858.

Implantable Cardioverter-Defibrillator (ICD) Therapy for Primary Prevention of Sudden Cardiac Death Might Not Provide Survival Benefit to Elderly

Clinical question: Does ICD therapy for primary prevention of sudden cardiac death in individuals with severe left ventricular dysfunction improve survival in elderly patients?

Background: Several clinical trials of select individuals with severe left ventricular dysfunction (EF <40%) have demonstrated that ICD therapy is associated with a reduction in overall mortality. Given the costs and risks associated with ICD placement, it is important to assess how this therapy affects survival in younger versus older individuals.

Study design: Meta-analysis of five randomized controlled trials.

Setting: Five trials (MADIT-II, DEFINITE, DINAMIT, SCDHeFT, and IRIS).

Synopsis: All included studies compared ICD therapy to standard medical care in the primary prevention of sudden cardiac death in individuals with severe left ventricular dysfunction. Elderly patients comprised 44% of the 5,783 patients included in the study; elderly patients were defined as >65 in three studies, and >60 in two studies. Pooled analysis of the three trials examining ICD use for primary prevention found a nonsignificant reduction in all-cause mortality compared with medical therapy in elderly patients (HR 0.81 [95% CI 0.62 to 1.05], P=0.11). The two remaining studies involving post-MI patients showed no statistically significant reduction in mortality in elderly patients.

In contrast, pooled data from younger patients in the five trials showed that prophylactic ICD therapy reduced all-cause mortality.

None of the selected trials demonstrated evidence of selection, performance, detection, or attrition bias. Only a small number of studies were included, and four potentially relevant trials were not included because mortality data by age group were not available.

Trials including cardiac resynchronization therapy (CRT) were excluded from this meta-analysis, and elderly patients meeting criteria for CRT comprise an important group that must be considered separately.

Bottom line: There is no definitive reduction in mortality with prophylactic ICD therapy in elderly patients with severe left ventricular dysfunction.

Citation: Santangeli P, Di Biase L, Dello Russo A, et al. Meta-analysis: age and effectiveness of prophylactic implantable cardioverter-defibrillators. Ann Intern Med. 2010;153(9):592-599.

System Delay Is Mortality Marker in STEMI Patients Treated with Primary PCI

Clinical question: What is the relationship between system delays to reperfusion therapy and mortality in patients with ST-segment-elevation myocardial infarction (STEMI)?

Background: In patients with STEMI, an early reperfusion strategy is often sought, and several studies have focused on the association between door-to-balloon delay and outcomes. Focusing more broadly on the time from first contact with the healthcare system to the initiation of reperfusion therapy (system delay) might be a more relevant approach.

Study design: Historical follow-up study.

Setting: Three high-volume PCI centers in Denmark.

Synopsis: Using population-based medical registries of 6,209 Danish patients treated for STEMI with percutaneous coronary intervention (PCI), the authors examined the association between delays in reperfusion therapy and mortality. System delay encompassed the entire time from first contact with the healthcare system to the initiation of reperfusion therapy.

Overall, shorter system delay was associated with significantly decreased mortality, as were the individual components of system delay (prehospital delay and door-to-balloon delay). The shortest delays (0-60 minutes) corresponded to a mortality rate of 15%, and the longest delays (181-360 minutes) corresponded to a mortality rate of 31%. Patients were followed for a median of 3.4 years.

This study is unique, as it is the first to look at the association between system delay and outcomes in patients with STEMI transported by EMS and treated with primary PCI. The study highlights the harmful impact of longer system delays on mortality. Limitations of this study include possible underestimation of system delay and challenges surrounding the right marker for reperfusion.

Bottom line: System delay might serve as a broad, comprehensive marker for predicting mortality in patients with STEMI treated with primary PCI.

Citation: Terkelsen CJ, Sorensen JT, Maeng M, et al. System delay and mortality among patients with STEMI treated with primary percutaneous coronary intervention. JAMA. 2010;304(7): 763-771.

D-Dimer Is Effective in Determining the Risk of VTE Recurrence after First Unprovoked Event

Clinical question: Does the timing, patient age, or cut point level affect the predictive value of a D-dimer in predicting VTE recurrence after a first unprovoked event?

Background: Anticoagulation duration after a first unprovoked VTE is at least three months, but it can be indefinite in patients with stable anticoagulation and low bleeding risk. Measuring a D-dimer level after discontinuation of anticoagulation is helpful in determining which patients might benefit from prolonged anticoagulation. However, several unanswered questions remain regarding D-dimer testing.

Study design: Patient-level meta-analysis.

Setting: Pooled patient-level data from seven prospective studies.

Synopsis: Patient-level data were obtained for all patients enrolled using post-treatment D-dimer measurement to predict recurrent VTE in patients with a first unprovoked VTE who had completed at least three months of anticoagulation therapy. The mean length of follow-up was 30 months. Patients with a positive D-dimer had recurrent VTE at a rate of 8.8 per 100 patient-years while those with a negative D-dimer had a rate of 3.7 per 100 patient-years.

Univariate analysis revealed an HR of 2.59 for patients with a positive versus a negative test result. The analysis also showed that the timing of the test, the age of the patient, and the actual cut points used for the various D-dimer tests did not affect the analysis significantly.

These studies’ strength is their large sample sizes and the use of prospective studies. The weaknesses include a mostly white patient population and incomplete data on all patients.

Bottom line: D-dimer testing is useful in predicting VTE recurrence after treatment for a first unprovoked event regardless of patient age, post-treatment timing, or the assay cut point used.

Citation: Douketis J, Tosetto A, Marcucci M, et al. Patient-level meta-analysis: effect of measurement timing, threshold, and patient age on ability of D-dimer testing to assess recurrence risk after unprovoked venous thromboembolism. Ann Intern Med. 2010;153(8): 523-531.

Thigh-High Stockings Are Better than Knee-High Stockings for Post-Stroke DVT Prophylaxis

Clinical question: Are thigh-high compression stockings better then knee-high stockings in immobilized acute-stroke patients?

Background: DVT is common in hospitalized stroke patients with immobility. Graduated compression stockings are often used for DVT prophylaxis, but the CLOTS-1 trial recently found that thigh-high stockings were ineffective after acute stroke. It is unclear if the more commonly used knee-high stockings are more effective than thigh-high stockings.

Study design: Parallel-group trial (the CLOTS-2 trial).

Setting: One hundred twelve hospitals in nine countries.

Synopsis: More than 3,100 patients with acute stroke and immobilization were recruited from January 2002 to May 2009. Patients were randomized to receive thigh-high or knee-high stockings. Patients also received usual care, including anticoagulants and a screening ultrasound for asymptomatic proximal DVT at seven to 10 days. Approximately 640 patients in each group also underwent ultrasound at 25-30 days.

Overall, 6.3% of patients in the thigh-high group had DVT, compared with 8.8% in the knee-high group (P=0.007). There were no significant differences in the secondary outcomes of pulmonary embolism or death. The thigh-high stockings had a higher number of adverse skin events. Enrollment was stopped early when the CLOTS-1 trial showed no difference in DVT rates between thigh-high stockings and no stockings.

Bottom line: Knee-high graduated compression stockings lead to worse outcomes than thigh-high stockings for DVT prophylaxis in immobilized acute-stroke patients.

Citation: CLOTS (Clots in Legs Or sTockings after Stroke) Trial Collaboration. Thigh-length versus below-knee stockings for deep venous thrombosis prophylaxis after stroke: a randomized trial. Ann Intern Med. 2010;153(9):553-562. TH

Pediatric HM Literature

Co-Infection with Pertussis in Infants Hospitalized for Bronchiolitis

Reviewed by Pediatric Editor Mark Shen, MD, medical director of hospital medicine at Dell Children’s Medical Center, Austin, Texas.

Clinical question: How often are infants admitted with bronchiolitis co-infected with Bordatella pertussis?

Background: Infants admitted for bronchiolitis have previously been reported to have co-infection with B. pertussis. Given the nonspecific symptoms associated with early B. pertussis infection, as well as the epidemiology of ongoing outbreaks, it might be useful to better define the risk of co-infection in the seasonal surge of acute bronchiolitis.

Study design: Retrospective cohort study.

Setting: One university hospital in Finland.

Synopsis: A prior study had generated nasopharyngeal aspirate samples for viral antigen detection from 205 healthy, full-term infants younger than 6 months hospitalized for bronchiolitis from 2001 to 2004. Of these samples, 142 (69%) were of quality sufficient for B. pertussis PCR testing in 2009.

Twelve (8.5%) of the 142 infants admitted with bronchiolitis were found to have B. pertussis. Eleven of the 12 infants were co-infected with at least one other virus (RSV in eight of them). Infants who tested positive for pertussis were more often found to have coughing spells (41.7% vs. 14.6% in those who tested negative).

The reported rate of co-infection in this study is higher than other recent reports from the U.S., likely due to the fact that samples were collected during a time period when the incidence of pertussis was relatively high in Finland.

Nevertheless, given the magnitude of a recent pertussis outbreak in California, it might be useful to consider co-infection in young infants admitted for bronchiolitis in areas with a relatively higher incidence of pertussis.

Although the retrospective nature of this study and the lack of a reported definition of coughing spells make further conclusions difficult to draw, cough might deserve further scrutiny during bronchiolitis season.

Bottom line: Consider co-infection with pertussis in young infants admitted with bronchiolitis.

Citation: Nuolivirta K, Koponen P, He Q, et al. Bordatella pertussis infection is common in nonvaccinated infants admitted for bronchiolitis. Pediatr Infect Dis J. 2010;29(11):1013-1015.

In This Edition

Literature at a Glance

A guide to this month’s studies

• Effects of extended VTE prophylaxis in medical patients
• Outcomes with and without preprocedural statins
• Association of subclinical hypothyroidism and CHD
• BP treatment after intracerebral hemorrhage
• Outcomes of ICD therapy in the elderly
• Systems delays and outcomes of STEMI
• D-dimer to predict recurrent VTE
• Stocking height and risk of post-stroke DVT

Extending Anticoagulant Prophylaxis after Medical Hospitalization Decreases VTE, Increases Major Bleeding

Clinical question: For patients with acute medical illness, does extending low-molecular-weight heparin (LMWH) administration for up to 28 days after discharge reduce the incidence of venous thromboembolism (VTE)?

Background: DVT and pulmonary embolism (PE) are common hospital-acquired complications. LMWH has been shown to reduce VTE for medical and surgical patients, and extended-duration LMWH reduces VTE in high-risk surgical patients. Whether extending anticoagulant prophylaxis after discharge for acutely ill medical patients with reduced mobility improves outcomes is unknown.

Study design: Randomized, placebo-controlled trial.

Setting: Three hundred seventy hospitals in 20 countries.

Synopsis: Eligible patients were >40 years old, hospitalized with acute medical illness, and had reduced mobility for ≥3 days. Patients received enoxaparin 40 mg SC daily prophylaxis while hospitalized and were then randomized to an additional 28±4 days of enoxaparin or placebo. Patients received a screening ultrasound to assess for asymptomatic DVT. The primary outcome was a composite of asymptomatic proximal DVT, symptomatic DVT or PE, or fatal PE during the period of extended prophylaxis.

An interim analysis indicated that extended prophylaxis was ineffective; at that time, the protocol was amended to target patients with severe immobility or with moderate immobility plus an additional risk factor (e.g. cancer, prior VTE, or age >75).

The study found that extended prophylaxis decreased the composite VTE outcome (2.5% vs 4.0%, P<0.05) and symptomatic VTE (0.2% vs 1.0%, P<0.05). The incidence of major bleeding was increased in the extended prophylaxis group (0.8% vs 0.3%, P<0.05). There was no difference in mortality.

The unplanned, midstudy protocol amendment to target higher-risk patients is a concern, though the final analyses included patients pre- and post-amendment.

Bottom line: Extending LMWH beyond hospitalization for patients admitted with acute medical illness and decreased mobility decreases VTE, but increases major bleeding.

Citation: Hull RD, Schellong SM, Tapson VF, et al. Extended-duration venous thromboembolism prophylaxis in acutely ill medical patients with recently reduced mobility: a randomized trial. Ann Intern Med. 2010;153(1);8-18.

Preprocedural Statin Therapy Reduces Postprocedural Myocardial Infarction

Clinical question: Does statin therapy reduce periprocedural cardiovascular events?

Background: Myocardial infarction (MI) and death are inherent risks of invasive procedures. Reduction of these risks in certain patient populations has been shown with the use of a beta blockade. Statins have shown promise during acute coronary syndrome. Questions remain about the role of statin therapy before invasive procedures in reducing adverse cardiovascular events.

Study design: Meta-analysis of randomized controlled trials.

Setting: Twenty-one studies involving 4,805 patients, published from inception of MEDLINE, Cochrane, and Clinicaltrials to February 2010.

Synopsis: The use of statins one to seven days preprocedure significantly reduced post-procedural MI in percutaneous coronary interventions (PCI) (P<0.0001). Statins given approximately four weeks in advance of noncardiac surgical procedures also significantly reduced postprocedural MI (P=0.004). An absolute risk reduction of 5.8% for postprocedural MI was found after PCI and 4.1% in noncardiac surgical procedures.

Statins did not show a significant reduction in postprocedural MI (P=0.40) or all-cause mortality (P=0.15) in coronary artery bypass graft surgery (CABG). However, statins did reduce post-CABG atrial fibrillation (P<0.0001).

The 21 studies used a variety of drugs and doses. However, the PCI studies favored atorvastatin 40 mg; more than half the CABG studies used atorvastatin 20 mg; and 91% of the noncardiac surgical studies used fluvastatin 80 mg. Dedicated trials are needed to demonstrate optimal statin agent, dose, and timing of therapy.

Bottom line: Preprocedural statin therapy reduces postprocedural MI after both PCI and noncardiac procedures but not after CABG.

Citation: Winchester DE, Wen X, Xie L, Bavry AA. Evidence of pre-procedural statin therapy: a meta-analysis of randomized trials. J Am Coll Cardiol. 2010;56(19); 1099-1109.

Subclinical Hypothyroidism Increases the Risk of Coronary Heart Disease and Mortality

Clinical question: What are the risks of coronary heart disease (CHD) and mortality among adults with subclinical hypothyroidism?

Background: Subclinical hypo-thyroidism is defined as an elevated serum thyroid stimulating hormone (TSH) level with a normal T4 concentration. Controversy exists regarding the treatment of subclinical hypothyroidism. Because of the association with hyperlipidemia and atherosclerosis, treatment of subclinical hypothyroidism is thought to be beneficial. Previous data from large prospective cohort studies regarding this association are conflicting.

Study design: Study-level meta-analysis of prospective cohort studies.

Setting: Eleven prospective cohorts in the U.S., Europe, Australia, Brazil, and Japan from 1972 to 2007.

Synopsis: Among 55,287 adults, 3,450 (6.2%) had subclinical hypothyroidism and 51,837 were euthyroid. Using Cox proportional hazard models, the association of subclinical hypothyroidism with CHD and mortality were determined for each cohort.

The risk of CHD events and CHD mortality increased with higher TSH concentrations.

In age- and sex-adjusted analyses, the hazard ratio (HR) for CHD events were as follows: HR=1.0 (TSH=4.5-6.9 mIU/L); HR=1.17 (TSH=7-9.9 mIU/L), and HR=1.89 (TSH=10-19.9 mIU/L). Similarly, HRs for CHD mortality showed an increasing trend: 1.09, 1.42, and 1.58, respectively.

Although the association is clearly established here, randomized controlled trials are needed to address whether thyroxine replacement can prevent CHD and the TSH threshold that will provide the most clinical benefit.

Bottom line: Subclinical hypo-thyroidism is associated with an increased risk for CHD events and mortality, primarily in patients with TSH concentrations of 10 mIU/L or higher.

Citation: Rodondi N, den Elzen WP, Bauer DC, et al. Subclinical hypothyroidism and the risk of coronary heart disease and mortality. JAMA. 2010;304(12): 1365-1374.

Reduction in Hematoma Growth after Acute Intracerebral Hemorrhage Associated with Lower Blood Pressure

Clinical question: Does intensive systolic blood pressure (SBP) <140 mmHg within one hour reduce hematoma growth after acute intracerebral bleeding?

Background: Early elevation of blood pressure after an acute intracerebral bleed is strongly associated with hematoma growth and worse outcomes. The pilot phase of the Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial (INTERACT) showed that early intensive blood pressure lowering reduced hematoma growth within six hours after onset of intracerebral hemorrhage.

Study design: Randomized controlled trial in which patients received an early intensive blood-pressure-lowering treatment (goal SBP<140 mmHg within one hour) or the AHA recommended best practice guidelines (goal SBP <180 mmHg), which were published in 1999.

Setting: Network of hospitals in China, South Korea, and Australia.

Synopsis: From 2005 to 2007, the study examined 404 patients with computed-tomography (CT) evidence of intracerebral hemorrhage, elevated SBP (150 mmHg-220 mmHg), and ability to commence BP lowering treatment within six hours of onset. Hematoma volumes were measured in the intensive treatment and guideline groups based on CT scans done at baseline and after 24 hours.

No significant association was found between the baseline SBP and the absolute or proportional growth of the hematoma. However, in the intensively treated patients who achieved target SBP within the first 24 hours, a significant association was found with the absolute and proportional hematoma growth.

Maximum reduction of hematoma growth occurred in the group with a median SBP of 135 mmHg.

This single-observational study did not include patients with severe intracranial bleeding who died or required surgical evacuation within the first 24 hours.

Hematoma size reduction did not improve survival or outcomes.

Bottom line: Intensive SBP lowering from 140 mmHg to 130 mmHg within one hour reduces hematoma growth after an intracranial hemorrhage.

Citation: Arima H, Anderson CS, Wang JG, et al. Lower treatment blood pressure is associated with the greatest reduction in hematoma growth after acute intracerebral hemorrhage. Hypertension. 2010;56(5):852-858.

Implantable Cardioverter-Defibrillator (ICD) Therapy for Primary Prevention of Sudden Cardiac Death Might Not Provide Survival Benefit to Elderly

Clinical question: Does ICD therapy for primary prevention of sudden cardiac death in individuals with severe left ventricular dysfunction improve survival in elderly patients?

Background: Several clinical trials of select individuals with severe left ventricular dysfunction (EF <40%) have demonstrated that ICD therapy is associated with a reduction in overall mortality. Given the costs and risks associated with ICD placement, it is important to assess how this therapy affects survival in younger versus older individuals.

Study design: Meta-analysis of five randomized controlled trials.

Setting: Five trials (MADIT-II, DEFINITE, DINAMIT, SCDHeFT, and IRIS).

Synopsis: All included studies compared ICD therapy to standard medical care in the primary prevention of sudden cardiac death in individuals with severe left ventricular dysfunction. Elderly patients comprised 44% of the 5,783 patients included in the study; elderly patients were defined as >65 in three studies, and >60 in two studies. Pooled analysis of the three trials examining ICD use for primary prevention found a nonsignificant reduction in all-cause mortality compared with medical therapy in elderly patients (HR 0.81 [95% CI 0.62 to 1.05], P=0.11). The two remaining studies involving post-MI patients showed no statistically significant reduction in mortality in elderly patients.

In contrast, pooled data from younger patients in the five trials showed that prophylactic ICD therapy reduced all-cause mortality.

None of the selected trials demonstrated evidence of selection, performance, detection, or attrition bias. Only a small number of studies were included, and four potentially relevant trials were not included because mortality data by age group were not available.

Trials including cardiac resynchronization therapy (CRT) were excluded from this meta-analysis, and elderly patients meeting criteria for CRT comprise an important group that must be considered separately.

Bottom line: There is no definitive reduction in mortality with prophylactic ICD therapy in elderly patients with severe left ventricular dysfunction.

Citation: Santangeli P, Di Biase L, Dello Russo A, et al. Meta-analysis: age and effectiveness of prophylactic implantable cardioverter-defibrillators. Ann Intern Med. 2010;153(9):592-599.

System Delay Is Mortality Marker in STEMI Patients Treated with Primary PCI

Clinical question: What is the relationship between system delays to reperfusion therapy and mortality in patients with ST-segment-elevation myocardial infarction (STEMI)?

Background: In patients with STEMI, an early reperfusion strategy is often sought, and several studies have focused on the association between door-to-balloon delay and outcomes. Focusing more broadly on the time from first contact with the healthcare system to the initiation of reperfusion therapy (system delay) might be a more relevant approach.

Study design: Historical follow-up study.

Setting: Three high-volume PCI centers in Denmark.

Synopsis: Using population-based medical registries of 6,209 Danish patients treated for STEMI with percutaneous coronary intervention (PCI), the authors examined the association between delays in reperfusion therapy and mortality. System delay encompassed the entire time from first contact with the healthcare system to the initiation of reperfusion therapy.

Overall, shorter system delay was associated with significantly decreased mortality, as were the individual components of system delay (prehospital delay and door-to-balloon delay). The shortest delays (0-60 minutes) corresponded to a mortality rate of 15%, and the longest delays (181-360 minutes) corresponded to a mortality rate of 31%. Patients were followed for a median of 3.4 years.

This study is unique, as it is the first to look at the association between system delay and outcomes in patients with STEMI transported by EMS and treated with primary PCI. The study highlights the harmful impact of longer system delays on mortality. Limitations of this study include possible underestimation of system delay and challenges surrounding the right marker for reperfusion.

Bottom line: System delay might serve as a broad, comprehensive marker for predicting mortality in patients with STEMI treated with primary PCI.

Citation: Terkelsen CJ, Sorensen JT, Maeng M, et al. System delay and mortality among patients with STEMI treated with primary percutaneous coronary intervention. JAMA. 2010;304(7): 763-771.

D-Dimer Is Effective in Determining the Risk of VTE Recurrence after First Unprovoked Event

Clinical question: Does the timing, patient age, or cut point level affect the predictive value of a D-dimer in predicting VTE recurrence after a first unprovoked event?

Background: Anticoagulation duration after a first unprovoked VTE is at least three months, but it can be indefinite in patients with stable anticoagulation and low bleeding risk. Measuring a D-dimer level after discontinuation of anticoagulation is helpful in determining which patients might benefit from prolonged anticoagulation. However, several unanswered questions remain regarding D-dimer testing.

Study design: Patient-level meta-analysis.

Setting: Pooled patient-level data from seven prospective studies.

Synopsis: Patient-level data were obtained for all patients enrolled using post-treatment D-dimer measurement to predict recurrent VTE in patients with a first unprovoked VTE who had completed at least three months of anticoagulation therapy. The mean length of follow-up was 30 months. Patients with a positive D-dimer had recurrent VTE at a rate of 8.8 per 100 patient-years while those with a negative D-dimer had a rate of 3.7 per 100 patient-years.

Univariate analysis revealed an HR of 2.59 for patients with a positive versus a negative test result. The analysis also showed that the timing of the test, the age of the patient, and the actual cut points used for the various D-dimer tests did not affect the analysis significantly.

These studies’ strength is their large sample sizes and the use of prospective studies. The weaknesses include a mostly white patient population and incomplete data on all patients.

Bottom line: D-dimer testing is useful in predicting VTE recurrence after treatment for a first unprovoked event regardless of patient age, post-treatment timing, or the assay cut point used.

Citation: Douketis J, Tosetto A, Marcucci M, et al. Patient-level meta-analysis: effect of measurement timing, threshold, and patient age on ability of D-dimer testing to assess recurrence risk after unprovoked venous thromboembolism. Ann Intern Med. 2010;153(8): 523-531.

Thigh-High Stockings Are Better than Knee-High Stockings for Post-Stroke DVT Prophylaxis

Clinical question: Are thigh-high compression stockings better then knee-high stockings in immobilized acute-stroke patients?

Background: DVT is common in hospitalized stroke patients with immobility. Graduated compression stockings are often used for DVT prophylaxis, but the CLOTS-1 trial recently found that thigh-high stockings were ineffective after acute stroke. It is unclear if the more commonly used knee-high stockings are more effective than thigh-high stockings.

Study design: Parallel-group trial (the CLOTS-2 trial).

Setting: One hundred twelve hospitals in nine countries.

Synopsis: More than 3,100 patients with acute stroke and immobilization were recruited from January 2002 to May 2009. Patients were randomized to receive thigh-high or knee-high stockings. Patients also received usual care, including anticoagulants and a screening ultrasound for asymptomatic proximal DVT at seven to 10 days. Approximately 640 patients in each group also underwent ultrasound at 25-30 days.

Overall, 6.3% of patients in the thigh-high group had DVT, compared with 8.8% in the knee-high group (P=0.007). There were no significant differences in the secondary outcomes of pulmonary embolism or death. The thigh-high stockings had a higher number of adverse skin events. Enrollment was stopped early when the CLOTS-1 trial showed no difference in DVT rates between thigh-high stockings and no stockings.

Bottom line: Knee-high graduated compression stockings lead to worse outcomes than thigh-high stockings for DVT prophylaxis in immobilized acute-stroke patients.

Citation: CLOTS (Clots in Legs Or sTockings after Stroke) Trial Collaboration. Thigh-length versus below-knee stockings for deep venous thrombosis prophylaxis after stroke: a randomized trial. Ann Intern Med. 2010;153(9):553-562. TH

Pediatric HM Literature

Co-Infection with Pertussis in Infants Hospitalized for Bronchiolitis

Reviewed by Pediatric Editor Mark Shen, MD, medical director of hospital medicine at Dell Children’s Medical Center, Austin, Texas.

Clinical question: How often are infants admitted with bronchiolitis co-infected with Bordatella pertussis?

Background: Infants admitted for bronchiolitis have previously been reported to have co-infection with B. pertussis. Given the nonspecific symptoms associated with early B. pertussis infection, as well as the epidemiology of ongoing outbreaks, it might be useful to better define the risk of co-infection in the seasonal surge of acute bronchiolitis.

Study design: Retrospective cohort study.

Setting: One university hospital in Finland.

Synopsis: A prior study had generated nasopharyngeal aspirate samples for viral antigen detection from 205 healthy, full-term infants younger than 6 months hospitalized for bronchiolitis from 2001 to 2004. Of these samples, 142 (69%) were of quality sufficient for B. pertussis PCR testing in 2009.

Twelve (8.5%) of the 142 infants admitted with bronchiolitis were found to have B. pertussis. Eleven of the 12 infants were co-infected with at least one other virus (RSV in eight of them). Infants who tested positive for pertussis were more often found to have coughing spells (41.7% vs. 14.6% in those who tested negative).

The reported rate of co-infection in this study is higher than other recent reports from the U.S., likely due to the fact that samples were collected during a time period when the incidence of pertussis was relatively high in Finland.

Nevertheless, given the magnitude of a recent pertussis outbreak in California, it might be useful to consider co-infection in young infants admitted for bronchiolitis in areas with a relatively higher incidence of pertussis.

Although the retrospective nature of this study and the lack of a reported definition of coughing spells make further conclusions difficult to draw, cough might deserve further scrutiny during bronchiolitis season.

Bottom line: Consider co-infection with pertussis in young infants admitted with bronchiolitis.

Citation: Nuolivirta K, Koponen P, He Q, et al. Bordatella pertussis infection is common in nonvaccinated infants admitted for bronchiolitis. Pediatr Infect Dis J. 2010;29(11):1013-1015.

In This Edition

Literature at a Glance

A guide to this month’s studies

• Effects of extended VTE prophylaxis in medical patients
• Outcomes with and without preprocedural statins
• Association of subclinical hypothyroidism and CHD
• BP treatment after intracerebral hemorrhage
• Outcomes of ICD therapy in the elderly
• Systems delays and outcomes of STEMI
• D-dimer to predict recurrent VTE
• Stocking height and risk of post-stroke DVT

Extending Anticoagulant Prophylaxis after Medical Hospitalization Decreases VTE, Increases Major Bleeding

Clinical question: For patients with acute medical illness, does extending low-molecular-weight heparin (LMWH) administration for up to 28 days after discharge reduce the incidence of venous thromboembolism (VTE)?

Background: DVT and pulmonary embolism (PE) are common hospital-acquired complications. LMWH has been shown to reduce VTE for medical and surgical patients, and extended-duration LMWH reduces VTE in high-risk surgical patients. Whether extending anticoagulant prophylaxis after discharge for acutely ill medical patients with reduced mobility improves outcomes is unknown.

Study design: Randomized, placebo-controlled trial.

Setting: Three hundred seventy hospitals in 20 countries.

Synopsis: Eligible patients were >40 years old, hospitalized with acute medical illness, and had reduced mobility for ≥3 days. Patients received enoxaparin 40 mg SC daily prophylaxis while hospitalized and were then randomized to an additional 28±4 days of enoxaparin or placebo. Patients received a screening ultrasound to assess for asymptomatic DVT. The primary outcome was a composite of asymptomatic proximal DVT, symptomatic DVT or PE, or fatal PE during the period of extended prophylaxis.

An interim analysis indicated that extended prophylaxis was ineffective; at that time, the protocol was amended to target patients with severe immobility or with moderate immobility plus an additional risk factor (e.g. cancer, prior VTE, or age >75).

The study found that extended prophylaxis decreased the composite VTE outcome (2.5% vs 4.0%, P<0.05) and symptomatic VTE (0.2% vs 1.0%, P<0.05). The incidence of major bleeding was increased in the extended prophylaxis group (0.8% vs 0.3%, P<0.05). There was no difference in mortality.

The unplanned, midstudy protocol amendment to target higher-risk patients is a concern, though the final analyses included patients pre- and post-amendment.

Bottom line: Extending LMWH beyond hospitalization for patients admitted with acute medical illness and decreased mobility decreases VTE, but increases major bleeding.

Citation: Hull RD, Schellong SM, Tapson VF, et al. Extended-duration venous thromboembolism prophylaxis in acutely ill medical patients with recently reduced mobility: a randomized trial. Ann Intern Med. 2010;153(1);8-18.

Preprocedural Statin Therapy Reduces Postprocedural Myocardial Infarction

Clinical question: Does statin therapy reduce periprocedural cardiovascular events?

Background: Myocardial infarction (MI) and death are inherent risks of invasive procedures. Reduction of these risks in certain patient populations has been shown with the use of a beta blockade. Statins have shown promise during acute coronary syndrome. Questions remain about the role of statin therapy before invasive procedures in reducing adverse cardiovascular events.

Study design: Meta-analysis of randomized controlled trials.

Setting: Twenty-one studies involving 4,805 patients, published from inception of MEDLINE, Cochrane, and Clinicaltrials to February 2010.

Synopsis: The use of statins one to seven days preprocedure significantly reduced post-procedural MI in percutaneous coronary interventions (PCI) (P<0.0001). Statins given approximately four weeks in advance of noncardiac surgical procedures also significantly reduced postprocedural MI (P=0.004). An absolute risk reduction of 5.8% for postprocedural MI was found after PCI and 4.1% in noncardiac surgical procedures.

Statins did not show a significant reduction in postprocedural MI (P=0.40) or all-cause mortality (P=0.15) in coronary artery bypass graft surgery (CABG). However, statins did reduce post-CABG atrial fibrillation (P<0.0001).

The 21 studies used a variety of drugs and doses. However, the PCI studies favored atorvastatin 40 mg; more than half the CABG studies used atorvastatin 20 mg; and 91% of the noncardiac surgical studies used fluvastatin 80 mg. Dedicated trials are needed to demonstrate optimal statin agent, dose, and timing of therapy.

Bottom line: Preprocedural statin therapy reduces postprocedural MI after both PCI and noncardiac procedures but not after CABG.

Citation: Winchester DE, Wen X, Xie L, Bavry AA. Evidence of pre-procedural statin therapy: a meta-analysis of randomized trials. J Am Coll Cardiol. 2010;56(19); 1099-1109.

Subclinical Hypothyroidism Increases the Risk of Coronary Heart Disease and Mortality

Clinical question: What are the risks of coronary heart disease (CHD) and mortality among adults with subclinical hypothyroidism?

Background: Subclinical hypo-thyroidism is defined as an elevated serum thyroid stimulating hormone (TSH) level with a normal T4 concentration. Controversy exists regarding the treatment of subclinical hypothyroidism. Because of the association with hyperlipidemia and atherosclerosis, treatment of subclinical hypothyroidism is thought to be beneficial. Previous data from large prospective cohort studies regarding this association are conflicting.

Study design: Study-level meta-analysis of prospective cohort studies.

Setting: Eleven prospective cohorts in the U.S., Europe, Australia, Brazil, and Japan from 1972 to 2007.

Synopsis: Among 55,287 adults, 3,450 (6.2%) had subclinical hypothyroidism and 51,837 were euthyroid. Using Cox proportional hazard models, the association of subclinical hypothyroidism with CHD and mortality were determined for each cohort.

The risk of CHD events and CHD mortality increased with higher TSH concentrations.

In age- and sex-adjusted analyses, the hazard ratio (HR) for CHD events were as follows: HR=1.0 (TSH=4.5-6.9 mIU/L); HR=1.17 (TSH=7-9.9 mIU/L), and HR=1.89 (TSH=10-19.9 mIU/L). Similarly, HRs for CHD mortality showed an increasing trend: 1.09, 1.42, and 1.58, respectively.

Although the association is clearly established here, randomized controlled trials are needed to address whether thyroxine replacement can prevent CHD and the TSH threshold that will provide the most clinical benefit.

Bottom line: Subclinical hypo-thyroidism is associated with an increased risk for CHD events and mortality, primarily in patients with TSH concentrations of 10 mIU/L or higher.

Citation: Rodondi N, den Elzen WP, Bauer DC, et al. Subclinical hypothyroidism and the risk of coronary heart disease and mortality. JAMA. 2010;304(12): 1365-1374.

Reduction in Hematoma Growth after Acute Intracerebral Hemorrhage Associated with Lower Blood Pressure

Clinical question: Does intensive systolic blood pressure (SBP) <140 mmHg within one hour reduce hematoma growth after acute intracerebral bleeding?

Background: Early elevation of blood pressure after an acute intracerebral bleed is strongly associated with hematoma growth and worse outcomes. The pilot phase of the Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial (INTERACT) showed that early intensive blood pressure lowering reduced hematoma growth within six hours after onset of intracerebral hemorrhage.

Study design: Randomized controlled trial in which patients received an early intensive blood-pressure-lowering treatment (goal SBP<140 mmHg within one hour) or the AHA recommended best practice guidelines (goal SBP <180 mmHg), which were published in 1999.

Setting: Network of hospitals in China, South Korea, and Australia.

Synopsis: From 2005 to 2007, the study examined 404 patients with computed-tomography (CT) evidence of intracerebral hemorrhage, elevated SBP (150 mmHg-220 mmHg), and ability to commence BP lowering treatment within six hours of onset. Hematoma volumes were measured in the intensive treatment and guideline groups based on CT scans done at baseline and after 24 hours.

No significant association was found between the baseline SBP and the absolute or proportional growth of the hematoma. However, in the intensively treated patients who achieved target SBP within the first 24 hours, a significant association was found with the absolute and proportional hematoma growth.

Maximum reduction of hematoma growth occurred in the group with a median SBP of 135 mmHg.

This single-observational study did not include patients with severe intracranial bleeding who died or required surgical evacuation within the first 24 hours.

Hematoma size reduction did not improve survival or outcomes.

Bottom line: Intensive SBP lowering from 140 mmHg to 130 mmHg within one hour reduces hematoma growth after an intracranial hemorrhage.

Citation: Arima H, Anderson CS, Wang JG, et al. Lower treatment blood pressure is associated with the greatest reduction in hematoma growth after acute intracerebral hemorrhage. Hypertension. 2010;56(5):852-858.

Implantable Cardioverter-Defibrillator (ICD) Therapy for Primary Prevention of Sudden Cardiac Death Might Not Provide Survival Benefit to Elderly

Clinical question: Does ICD therapy for primary prevention of sudden cardiac death in individuals with severe left ventricular dysfunction improve survival in elderly patients?

Background: Several clinical trials of select individuals with severe left ventricular dysfunction (EF <40%) have demonstrated that ICD therapy is associated with a reduction in overall mortality. Given the costs and risks associated with ICD placement, it is important to assess how this therapy affects survival in younger versus older individuals.

Study design: Meta-analysis of five randomized controlled trials.

Setting: Five trials (MADIT-II, DEFINITE, DINAMIT, SCDHeFT, and IRIS).

Synopsis: All included studies compared ICD therapy to standard medical care in the primary prevention of sudden cardiac death in individuals with severe left ventricular dysfunction. Elderly patients comprised 44% of the 5,783 patients included in the study; elderly patients were defined as >65 in three studies, and >60 in two studies. Pooled analysis of the three trials examining ICD use for primary prevention found a nonsignificant reduction in all-cause mortality compared with medical therapy in elderly patients (HR 0.81 [95% CI 0.62 to 1.05], P=0.11). The two remaining studies involving post-MI patients showed no statistically significant reduction in mortality in elderly patients.

In contrast, pooled data from younger patients in the five trials showed that prophylactic ICD therapy reduced all-cause mortality.

None of the selected trials demonstrated evidence of selection, performance, detection, or attrition bias. Only a small number of studies were included, and four potentially relevant trials were not included because mortality data by age group were not available.

Trials including cardiac resynchronization therapy (CRT) were excluded from this meta-analysis, and elderly patients meeting criteria for CRT comprise an important group that must be considered separately.

Bottom line: There is no definitive reduction in mortality with prophylactic ICD therapy in elderly patients with severe left ventricular dysfunction.

Citation: Santangeli P, Di Biase L, Dello Russo A, et al. Meta-analysis: age and effectiveness of prophylactic implantable cardioverter-defibrillators. Ann Intern Med. 2010;153(9):592-599.

System Delay Is Mortality Marker in STEMI Patients Treated with Primary PCI

Clinical question: What is the relationship between system delays to reperfusion therapy and mortality in patients with ST-segment-elevation myocardial infarction (STEMI)?

Background: In patients with STEMI, an early reperfusion strategy is often sought, and several studies have focused on the association between door-to-balloon delay and outcomes. Focusing more broadly on the time from first contact with the healthcare system to the initiation of reperfusion therapy (system delay) might be a more relevant approach.

Study design: Historical follow-up study.

Setting: Three high-volume PCI centers in Denmark.

Synopsis: Using population-based medical registries of 6,209 Danish patients treated for STEMI with percutaneous coronary intervention (PCI), the authors examined the association between delays in reperfusion therapy and mortality. System delay encompassed the entire time from first contact with the healthcare system to the initiation of reperfusion therapy.

Overall, shorter system delay was associated with significantly decreased mortality, as were the individual components of system delay (prehospital delay and door-to-balloon delay). The shortest delays (0-60 minutes) corresponded to a mortality rate of 15%, and the longest delays (181-360 minutes) corresponded to a mortality rate of 31%. Patients were followed for a median of 3.4 years.

This study is unique, as it is the first to look at the association between system delay and outcomes in patients with STEMI transported by EMS and treated with primary PCI. The study highlights the harmful impact of longer system delays on mortality. Limitations of this study include possible underestimation of system delay and challenges surrounding the right marker for reperfusion.

Bottom line: System delay might serve as a broad, comprehensive marker for predicting mortality in patients with STEMI treated with primary PCI.

Citation: Terkelsen CJ, Sorensen JT, Maeng M, et al. System delay and mortality among patients with STEMI treated with primary percutaneous coronary intervention. JAMA. 2010;304(7): 763-771.

D-Dimer Is Effective in Determining the Risk of VTE Recurrence after First Unprovoked Event

Clinical question: Does the timing, patient age, or cut point level affect the predictive value of a D-dimer in predicting VTE recurrence after a first unprovoked event?

Background: Anticoagulation duration after a first unprovoked VTE is at least three months, but it can be indefinite in patients with stable anticoagulation and low bleeding risk. Measuring a D-dimer level after discontinuation of anticoagulation is helpful in determining which patients might benefit from prolonged anticoagulation. However, several unanswered questions remain regarding D-dimer testing.

Study design: Patient-level meta-analysis.

Setting: Pooled patient-level data from seven prospective studies.

Synopsis: Patient-level data were obtained for all patients enrolled using post-treatment D-dimer measurement to predict recurrent VTE in patients with a first unprovoked VTE who had completed at least three months of anticoagulation therapy. The mean length of follow-up was 30 months. Patients with a positive D-dimer had recurrent VTE at a rate of 8.8 per 100 patient-years while those with a negative D-dimer had a rate of 3.7 per 100 patient-years.

Univariate analysis revealed an HR of 2.59 for patients with a positive versus a negative test result. The analysis also showed that the timing of the test, the age of the patient, and the actual cut points used for the various D-dimer tests did not affect the analysis significantly.

These studies’ strength is their large sample sizes and the use of prospective studies. The weaknesses include a mostly white patient population and incomplete data on all patients.

Bottom line: D-dimer testing is useful in predicting VTE recurrence after treatment for a first unprovoked event regardless of patient age, post-treatment timing, or the assay cut point used.

Citation: Douketis J, Tosetto A, Marcucci M, et al. Patient-level meta-analysis: effect of measurement timing, threshold, and patient age on ability of D-dimer testing to assess recurrence risk after unprovoked venous thromboembolism. Ann Intern Med. 2010;153(8): 523-531.

Thigh-High Stockings Are Better than Knee-High Stockings for Post-Stroke DVT Prophylaxis

Clinical question: Are thigh-high compression stockings better then knee-high stockings in immobilized acute-stroke patients?

Background: DVT is common in hospitalized stroke patients with immobility. Graduated compression stockings are often used for DVT prophylaxis, but the CLOTS-1 trial recently found that thigh-high stockings were ineffective after acute stroke. It is unclear if the more commonly used knee-high stockings are more effective than thigh-high stockings.

Study design: Parallel-group trial (the CLOTS-2 trial).

Setting: One hundred twelve hospitals in nine countries.

Synopsis: More than 3,100 patients with acute stroke and immobilization were recruited from January 2002 to May 2009. Patients were randomized to receive thigh-high or knee-high stockings. Patients also received usual care, including anticoagulants and a screening ultrasound for asymptomatic proximal DVT at seven to 10 days. Approximately 640 patients in each group also underwent ultrasound at 25-30 days.

Overall, 6.3% of patients in the thigh-high group had DVT, compared with 8.8% in the knee-high group (P=0.007). There were no significant differences in the secondary outcomes of pulmonary embolism or death. The thigh-high stockings had a higher number of adverse skin events. Enrollment was stopped early when the CLOTS-1 trial showed no difference in DVT rates between thigh-high stockings and no stockings.

Bottom line: Knee-high graduated compression stockings lead to worse outcomes than thigh-high stockings for DVT prophylaxis in immobilized acute-stroke patients.

Citation: CLOTS (Clots in Legs Or sTockings after Stroke) Trial Collaboration. Thigh-length versus below-knee stockings for deep venous thrombosis prophylaxis after stroke: a randomized trial. Ann Intern Med. 2010;153(9):553-562. TH

Pediatric HM Literature

Co-Infection with Pertussis in Infants Hospitalized for Bronchiolitis

Reviewed by Pediatric Editor Mark Shen, MD, medical director of hospital medicine at Dell Children’s Medical Center, Austin, Texas.

Clinical question: How often are infants admitted with bronchiolitis co-infected with Bordatella pertussis?

Background: Infants admitted for bronchiolitis have previously been reported to have co-infection with B. pertussis. Given the nonspecific symptoms associated with early B. pertussis infection, as well as the epidemiology of ongoing outbreaks, it might be useful to better define the risk of co-infection in the seasonal surge of acute bronchiolitis.

Study design: Retrospective cohort study.

Setting: One university hospital in Finland.

Synopsis: A prior study had generated nasopharyngeal aspirate samples for viral antigen detection from 205 healthy, full-term infants younger than 6 months hospitalized for bronchiolitis from 2001 to 2004. Of these samples, 142 (69%) were of quality sufficient for B. pertussis PCR testing in 2009.

Twelve (8.5%) of the 142 infants admitted with bronchiolitis were found to have B. pertussis. Eleven of the 12 infants were co-infected with at least one other virus (RSV in eight of them). Infants who tested positive for pertussis were more often found to have coughing spells (41.7% vs. 14.6% in those who tested negative).

The reported rate of co-infection in this study is higher than other recent reports from the U.S., likely due to the fact that samples were collected during a time period when the incidence of pertussis was relatively high in Finland.

Nevertheless, given the magnitude of a recent pertussis outbreak in California, it might be useful to consider co-infection in young infants admitted for bronchiolitis in areas with a relatively higher incidence of pertussis.

Although the retrospective nature of this study and the lack of a reported definition of coughing spells make further conclusions difficult to draw, cough might deserve further scrutiny during bronchiolitis season.

Bottom line: Consider co-infection with pertussis in young infants admitted with bronchiolitis.

Citation: Nuolivirta K, Koponen P, He Q, et al. Bordatella pertussis infection is common in nonvaccinated infants admitted for bronchiolitis. Pediatr Infect Dis J. 2010;29(11):1013-1015.

In This Edition

Literature at a Glance

A guide to this month’s studies

• Perioperative oxygen use and infection rate.
• Effects of adverse events and healthcare costs
• Impact of DVT on PE rates in trauma
• VTE prevention and treatment in cancer patients
• Effect of perioperative beta-blocker discontinuation
• Endovascular vs. open AAA repair
• OTC analgesics in patients with hepatic dysfunction
• Cardiovascular disease and risk of hip fracture

High Perioperative Oxygen Fraction Does Not Improve Surgical-Site Infection Frequency after Abdominal Surgery

Clinical question: Does the use of 80% oxygen perioperatively in abdominal surgery decrease the frequency of surgical-site infection within 14 days without increasing the rate of pulmonary complications?

Background: Low oxygen tension in wounds can negatively impact immune response and healing. Increasing inspiratory oxygen fraction during the perioperative period translates into higher wound oxygen tension. However, the benefit of increased oxygen fraction therapy in abdominal surgery healing and complications is not clear, nor is the frequency of pulmonary complications.

Study design: Patient- and observer-blinded clinical trial.

Setting: Fourteen Danish hospitals from October 2006 to October 2008.

Synopsis: Patients were randomized to receive a fraction of inspired oxygen (FIO₂) of 0.80 or 0.30. The primary outcome—surgical-site infection in the superficial or deep wound or intra-abdominal cavity within 14 days of surgery—was defined using Centers for Disease Control and Prevention (CDC) criteria. Secondary outcomes included pulmonary complications within 14 days (pneumonia, atelectasis, or respiratory failure), 30-day mortality, duration of post-op course, ICU stay within 14 days post-op, and any abdominal operation within 14 days. The 1,386 patients were enrolled in the intention-to-treat analysis.

Infection occurred in 19.1% of patients given 0.80 FIO2 and in 20.1% of patients given 0.30 FIO2; odds ratio of 0.94 (95% CI 0.72 to 1.22; P=0.64). Numbers of pulmonary complications were not significantly different between the groups.

This trial included acute and nonacute laparotomies with followup for adverse outcomes. Study limitations included the inability to ensure that both groups received timely antibiotics and prevention for hypothermia. Of patients in the 30% FIO2 group, 7.3% required higher oxygen administration. Additionally, infection might have been underestimated in 11.3% of patients who were not followed up on between days 13 and 30.

Bottom line: High oxygen concentration administered during and after laparotomy did not lead to fewer surgical site infections, nor did it significantly increase the frequency of pulmonary complications or death.

Citation: Meyhoff CS, Wetterslev J, Jorgensen LN, et al. Effect of high perioperative oxygen fraction on surgical site infection and pulmonary complications after abdominal surgery: the PROXI randomized clinical trial. JAMA. 2009;302(14):1543-1550.

Eliminating Adverse Events and Redundant Tests Could Generate U.S. Healthcare Savings

Clinical question: Using available data, what is the estimated cost savings of eliminating adverse events and avoiding redundant tests?

Background: Reimbursement schemes are changing such that hospitals are reimbursed less for some adverse events. This financial disincentive is expected to spark interest in improved patient safety. The authors sought to model the cost savings generated by eliminating redundant testing and adverse events from literature-based estimates.

Study design: Development of conceptual model to identify common or costly adverse events, redundant tests, and simulated costs.

Setting: Literature review, expert opinion, data from safety organizations and epidemiologic studies, and patient data from the 2004 National Inpatient Data Sample.

Synopsis: The conceptual model identified 5.7 million adverse events in U.S. hospitals, of which 3 million were considered preventable. The most common events included hospital-acquired infections (82% preventable), adverse drug events (26%), falls (33%), and iatrogenic thromboembolic events (62%). The calculated cost savings totaled $16.6 billion (5.5% of total inpatient costs) for adverse events and $8.2 billion for the elimination of redundant tests. When looking at hospital subtypes, the greatest savings would come from major teaching hospitals.

This study is limited by its use of published and heterogeneous data spanning a 15-year period. The authors did not include events for which there was no epidemiologic or cost data. As hospital-care changes and technology is adopted, it is uncertain how this changes the costs, prevalence, and the preventable nature of these events. The model was not consistently able to identifying high- and low-risk patients. For instance, in some models, all patients were considered at risk for events.

Bottom line: Based on a conceptual model of 2004 hospitalized patients, eliminating preventable adverse events could have saved $16.6 billion, while eliminating redundant tests could have saved another $8 billion.

Citation: Jha AK, Chan DC, Ridgway AB, Franz C, Bates DW. Improving safety and eliminating redundant tests: cutting costs in U.S. hospitals. Health Aff (Millwood). 2009;28(5):1475-1484.

Trauma Patients with Pulmonary Embolism Might Not Have DVT on Imaging of Lower Extremities

Clinical question: What is the relationship between acute DVT and pulmonary embolism (PE) in trauma patients?

Background: Major trauma is associated with an increased risk of acute DVT and PE. It is assumed that the majority of PEs arise from DVTs in the lower extremities. Definitive evidence demonstrating that PEs form in situ rather than embolize from leg veins could impact indications for inferior vena cava filters.

Study design: Retrospective chart review.

Setting: Academic Level 1 trauma center in Boston.

Synopsis: The medical records of 247 trauma patients with suspected PE who underwent CT angiography of the lungs and simultaneous CT venography of the pelvis and lower extremities from January 2004 to December 2007 were reviewed. High-risk patients also underwent weekly screening with duplex ultrasonagraphy of the legs.

PE was diagnosed in 46 patients (19%) and DVT in 18 patients (7%). Anticoagulant prophylaxis had been administered to 96% and 78% of the patients with PE and DVT, respectively. PE was diagnosed a median of 5.5 days after admission (range 0-40 days) and the majority (61%) were in segmental or subsegmental branches, rather than in the main or lobar pulmonary arteries (39%). Only seven of the 46 patients (15%) diagnosed with PE also had a pelvic or lower-extremity DVT on simultaneous imaging with CT venography.

Bottom line: Trauma patients with PE often do not have a DVT at the time of diagnosis, though it remains unknown whether this is due to in-situ pulmonary thrombosis or complete embolization from the lower extremities.

Citation: Velmahos GC, Spaniolas K, Tabbara M, et al. Pulmonary embolism and deep venous thrombosis in trauma: are they related? Arch Surg. 2009;144:928-932.

Cancer Guideline for VTE Prophylaxis for Inpatients and Long-Term Treatment With Low-Molecular-Weight Heparin for Acute VTE

Clinical question: On what aspects of VTE management in cancer patients are there consensus among the major guideline panels?

Background: VTE is a common and serious complication of cancer. Patients might be hypercoagulable due to prothrombotic mediators released or mediated by tumor cells, chemotherapeutic agents, debility, central venous catheters, hospitalizations, or surgical procedures. The optimal management often is problematic due to uncertain benefit and risk of bleeding.

Study design: Review of major guideline statements.

Synopsis: The authors examined five VTE guidelines of American and European cancer societies. Each guideline was reviewed to determine the main recommendations and whether there was consensus on key aspects of anticoagulant management.

The study authors concluded that consensus was reached on most key recommendations:

• VTE prophylaxis in hospitalized medical patients. All five guidelines recommend the use of prophylaxis, though some guidelines recommend anticoagulant prophylaxis for all inpatients in the absence of contraindications and some recommend limiting prophylaxis to immobilized patients. All five recommend the use of either unfractionated heparin, low-molecular-weight heparin (LMWH), or fondaparinux.
• VTE prevention in cancer patients undergoing surgery. All five guidelines recommend anticoagulant prophylaxis in the absence of contraindications and extending prophylaxis approximately four weeks after major surgery.
• VTE prophylaxis in cancer patients with central venous catheters. Not recommended.
• VTE prophylaxis in ambulatory cancer patients without central venous catheters. Recommended only for multiple myeloma patients receiving a thalidomide-lenalidomide regimen.
• Long-term treatment of acute VTE in cancer patients. All five guidelines recommend initial treatment with LMWH for at least three to six months, followed by indefinite treatment with LMWH or a vitamin K antagonist.

Bottom line: Major guideline panels agree on key aspects of VTE management for cancer patients, including the use of prophylaxis for hospitalized medical and surgical patients and the use of long-term LMWH treatment for cancer patients with acute VTE.

Citation: Khorana AA, Streiff MB, Farge D, et al. Venous thromboembolism prophylaxis and treatment in cancer: a consensus statement of major guidelines panels and call to action. J Clin Oncol. 2009; 27(29):4919-4926.

Discontinuation of Beta Blockers Increases Risk of Postoperative Myocardial Infarction and Death

Clinical question: Does perioperative beta-blocker discontinuation affect postoperative myocardial infarction (MI) in low-risk patients undergoing joint arthroplasty?

Background: Recent trials show no benefit of perioperative beta blockers in reducing the incidence of perioperative myocardial infarctions (POMI) in low-risk patients. This retrospective study examined the impact of continuing or discontinuing beta blockers and the occurrence of POMI in patients undergoing elective joint arthroplasties.

Study design: Retrospective chart review.

Setting: Large academic center in Ottawa, Canada.

Synopsis: Medical records for 5,178 patients undergoing elective hip or knee arthroplasty from January 2002 to June 2006 were included in the review. The primary outcome was POMI, defined as an increased troponin level. Patients were divided into three groups: beta blocker prescribed on post-operative day (POD) zero and continued for one week or until discharge; beta blocker prescribed on POD zero and discontinued at any time in the first week; and no beta blocker on POD 0.

Beta blockers were continued in 992 patients and discontinued in 252 patients. The rate of POMI and death increased in the beta-blocker discontinuation group (odds ratio 2.0 [1.1-3.9] and 2.0 [1.1-3.9], respectively). This association persisted after adjustment for cardiac risk using a validated risk score.

The study was limited by the fact that the control group did not include patients who were on a beta blocker at home, thus potentially increasing the number of events in this group. The discontinuation beta blocker group had an increased baseline risk for POMI. The reason for discontinuing the beta blocker was not known, and cessation of beta blocker could have been due to an acute event.

Bottom line: This study adds support to the American College of Cardiology and American Heart Association (ACC/AHA) guidelines, which recommend continuation of beta-blocker therapy in the perioperative period.

Citation: Van Klei WA, Bryson GL, Yang H, Forster AJ. Effect of beta-blocker prescription on the incidence of postoperative myocardial infarction after hip and knee arthroplasty. Anesthesiology. 2009;111(4):717-724.

Lower Perioperative Mortality with Endovascular Vs. Open Abdominal Aortic Aneurysm Repair

Clinical question: How do perioperative and long-term morbidity and mortality compare in endovascular and open repair of abdominal aortic aneurysm (AAA)?

Background: Open AAA repair has relatively high perioperative mortality. Endovascular repair was developed as a less-invasive option and has been shown to reduce inpatient perioperative mortality, length of hospital stay, and ICU requirement. However, data suggest it leads to more frequent reinterventions and the same mortality rate as open repair at two years.

Study design: Randomized clinical trial.

Setting: Veterans Affairs medical centers.

Synopsis: The study randomized 881 veterans who planned to have elective AAA repair and were eligible for both endovascular and open repair. This is a planned, two-year interim report in a nine-year study.

Perioperative mortality was 0.5% in the endovascular repair group, compared with 3.0% in the open repair group. However, this difference in mortality was not statistically significant at two years. The endovascular repair group experienced shorter procedure and mechanical ventilation time, decreased hospital and ICU stay, and lower rate of blood transfusions.

Overall, there was no difference between the groups for major morbidity, procedure failure, need for secondary therapeutic intervention, quality of life, or erectile dysfunction. More data on long-term comparison of these two interventions will be available at the conclusion of this study.

Bottom line: Endovascular repair of AAA has lower perioperative mortality than open repair but did not lead to improved morbidity or mortality at two years.

Citation: Lederle FA, Freischlag JA, Kyriakides TC, et al. Outcomes following endovascular vs. open repair of abdominal aortic aneurysm: a randomized trial. JAMA. 2009;302 (14):1535-1542.

OTC Analgesics Not Associated with Acute Decompensation in Cirrhotic Patients

Clinical question: Do over-the-counter (OTC) analgesics lead to acute hepatic decompensation among patients with cirrhosis?

Background: In theory, intake of acetaminophen and/or nonsteroidal anti-inflammatory drugs (NSAIDs) can worsen hepatic function and lead to complications among cirrhotic patients. The role of OTC analgesics in potentially triggering acute hepatic decompensation among cirrhotic patients has not been studied.

Study design: Prospective case-control study.

Setting: Two tertiary-care hospitals.

Synopsis: Cirrhotic patients hospitalized for acute liver decompensation were compared with compensated cirrhotic patients in the liver clinic (cirrhotic controls) and with randomly selected, noncirrhotic patients who were simultaneously hospitalized (noncirrhotic controls). Data collected through questionnaires included quantity and dose of OTC analgesics used and alcohol consumption in the past 30 days.

Thirty-five percent of the hospitalized cirrhotic patients, 52% of the cirrhotic controls, and 70% of the noncirrhotic controls used OTC analgesics. At doses lower than those recommended, acetaminophen is not associated with acute liver decompensation among cirrhotic patients, even with recent alcohol use. However, NSAIDs taken by the cirrhotic patients, when compared to control subjects, were in larger doses and used for a longer duration, suggesting NSAIDs may have contributed to the acute decompensation.

Study limitations include the nature of the study design, reliance on the patient’s recall of OTC analgesic use, and obtaining other possible causes of decompensation, such as herbal supplement intake or compliance with diuretics or dietary indiscretion.

Bottom line: Acetaminophen at doses lower than recommended is not associated with adverse complications in cirrhotic patients, but NSAIDs are possibly associated with acute decompensation.

Citation: Khalid SK, Lane J, Navarro V, Garcia-Tsao G. Use of over-the-counter analgesics is not associated with acute decompensation in patients with cirrhosis. Clin Gastroenterol Hepatol. 2009;7(9):994-999.

Cardiovascular Disease and Risk of Hip Fracture

Clinical question: Is the diagnosis of cardiovascular disease (CVD) associated with the risk of subsequent hip fracture?

Background: Osteoporosis and CVD are regarded as independent, age-related conditions. However, recent research suggests that the bone and vascular systems share common regulatory mechanisms. Stroke is a known risk factor for hip fractures, and bisphosphonates have been shown to prevent atherosclerosis and reduce total mortality rate.

Study design: Cohort study.

Setting: Swedish National Patient Registry.

Synopsis: The study identified 31,936 Swedish twins born from 1914 to 1944. This cohort was followed up to age 50, and time-dependent exposures using Cox-proportional hazard regression models were evaluated.

Times to hip fracture after CVD diagnosis were isolated. Crude absolute rate of hip fractures (per 1,000 person-years) was 12.6 after diagnosis of heart failure, 12.6 after a stroke, 6.6 after peripheral atherosclerosis, and 5.2 after ischemic heart disease (IHD), compared with 1.2 per 1,000 person-years without a CVD diagnosis. Multivariable-adjusted hazard ratio (HR) of hip fracture after heart failure was 4.40 (95% CI, 3.43-5.63); after a stroke was 5.09 (95% CI, 4.18-6.20); after peripheral atherosclerosis was 3.20 (CI, 2.28-4.50); and after an IHD event was 2.32 (CI, 1.91-2.84).

Identical twins even without heart failure and stroke also had an increased risk of hip fracture if their twin had been diagnosed with these diseases.

Bottom line: Cardiovascular disease is significantly associated with risk of subsequent hip fracture, and genetic factors probably play a role in the association.

Citation: Sennerby U, Melhus H, Gedeborg R, et al. Cardiovascular diseases and risk of hip fracture. JAMA. 2009;302(15):1666-1673. TH

PEDIATRIC HM Literature

By Mark Shen, MD

Variation in the Treatment of Henoch-Schönlein Purpura

Reviewed by Pediatric Editor Mark Shen, MD, medical director of hospital medicine at Dell Children’s Medical Center, Austin, Texas.

Clinical question: What is the degree of variation in the inpatient management of Henoch-Schönlein purpura (HSP)?

Background: HSP is the most common pediatric vasculitis, but there are no consensus recommendations or guidelines for treatment. The amount of variation in the pharmacologic management of this disease is unknown.

Study design: Retrospective database analysis.

Setting: Thirty-six children’s hospitals affiliated with the Child Health Corporation of America.

Synopsis: The Pediatric Health Information (PHIS) database was sampled for children younger than 18 years of age with an ICD-9-CM code of HSP and discharge from a hospital that submitted appropriate data from 2000 to 2007. Only index admissions were included, and children with coexisting rheumatic conditions were excluded, for a total of 1,988 subjects.

Logistic regression analysis was used to examine the effects of patient-level standardization on hospital-level rates of therapy and the degree to which variation across hospitals occurred beyond what would be expected after standardization.

Hospital-level variation in medication use was significant (P<0.001) for corticosteroids, opiates, and nonsteroidal anti-inflammatory drugs (NSAIDs), even after adjustment for severity and age at presentation.

Although variation in management is not surprising, the significant degree to which this occurred at the hospital level suggests that local institutional culture plays a dominant role in decision-making. The use of the PHIS database allows for analysis of a large population that would be otherwise difficult to study. However, significant numbers of HSP patients do not require hospitalization, and the study results might substantially over- or underestimate practice patterns. Collaborative efforts to better define optimal management of HSP are needed.

Bottom line: A significant degree of hospital-level variation exists in the inpatient management of HSP.

Citation: Weiss PF, Klink AJ, Hexem K, et al. Variation in inpatient therapy and diagnostic evaluation of children with henoch schönlein purpura. J Pediatr. 2009;155(6):812-818.e1.

In This Edition

Literature at a Glance

A guide to this month’s studies

• Perioperative oxygen use and infection rate.
• Effects of adverse events and healthcare costs
• Impact of DVT on PE rates in trauma
• VTE prevention and treatment in cancer patients
• Effect of perioperative beta-blocker discontinuation
• Endovascular vs. open AAA repair
• OTC analgesics in patients with hepatic dysfunction
• Cardiovascular disease and risk of hip fracture

High Perioperative Oxygen Fraction Does Not Improve Surgical-Site Infection Frequency after Abdominal Surgery

Clinical question: Does the use of 80% oxygen perioperatively in abdominal surgery decrease the frequency of surgical-site infection within 14 days without increasing the rate of pulmonary complications?

Background: Low oxygen tension in wounds can negatively impact immune response and healing. Increasing inspiratory oxygen fraction during the perioperative period translates into higher wound oxygen tension. However, the benefit of increased oxygen fraction therapy in abdominal surgery healing and complications is not clear, nor is the frequency of pulmonary complications.

Study design: Patient- and observer-blinded clinical trial.

Setting: Fourteen Danish hospitals from October 2006 to October 2008.

Synopsis: Patients were randomized to receive a fraction of inspired oxygen (FIO₂) of 0.80 or 0.30. The primary outcome—surgical-site infection in the superficial or deep wound or intra-abdominal cavity within 14 days of surgery—was defined using Centers for Disease Control and Prevention (CDC) criteria. Secondary outcomes included pulmonary complications within 14 days (pneumonia, atelectasis, or respiratory failure), 30-day mortality, duration of post-op course, ICU stay within 14 days post-op, and any abdominal operation within 14 days. The 1,386 patients were enrolled in the intention-to-treat analysis.

Infection occurred in 19.1% of patients given 0.80 FIO2 and in 20.1% of patients given 0.30 FIO2; odds ratio of 0.94 (95% CI 0.72 to 1.22; P=0.64). Numbers of pulmonary complications were not significantly different between the groups.

This trial included acute and nonacute laparotomies with followup for adverse outcomes. Study limitations included the inability to ensure that both groups received timely antibiotics and prevention for hypothermia. Of patients in the 30% FIO2 group, 7.3% required higher oxygen administration. Additionally, infection might have been underestimated in 11.3% of patients who were not followed up on between days 13 and 30.

Bottom line: High oxygen concentration administered during and after laparotomy did not lead to fewer surgical site infections, nor did it significantly increase the frequency of pulmonary complications or death.

Citation: Meyhoff CS, Wetterslev J, Jorgensen LN, et al. Effect of high perioperative oxygen fraction on surgical site infection and pulmonary complications after abdominal surgery: the PROXI randomized clinical trial. JAMA. 2009;302(14):1543-1550.

Eliminating Adverse Events and Redundant Tests Could Generate U.S. Healthcare Savings

Clinical question: Using available data, what is the estimated cost savings of eliminating adverse events and avoiding redundant tests?

Background: Reimbursement schemes are changing such that hospitals are reimbursed less for some adverse events. This financial disincentive is expected to spark interest in improved patient safety. The authors sought to model the cost savings generated by eliminating redundant testing and adverse events from literature-based estimates.

Study design: Development of conceptual model to identify common or costly adverse events, redundant tests, and simulated costs.

Setting: Literature review, expert opinion, data from safety organizations and epidemiologic studies, and patient data from the 2004 National Inpatient Data Sample.

Synopsis: The conceptual model identified 5.7 million adverse events in U.S. hospitals, of which 3 million were considered preventable. The most common events included hospital-acquired infections (82% preventable), adverse drug events (26%), falls (33%), and iatrogenic thromboembolic events (62%). The calculated cost savings totaled $16.6 billion (5.5% of total inpatient costs) for adverse events and $8.2 billion for the elimination of redundant tests. When looking at hospital subtypes, the greatest savings would come from major teaching hospitals.

This study is limited by its use of published and heterogeneous data spanning a 15-year period. The authors did not include events for which there was no epidemiologic or cost data. As hospital-care changes and technology is adopted, it is uncertain how this changes the costs, prevalence, and the preventable nature of these events. The model was not consistently able to identifying high- and low-risk patients. For instance, in some models, all patients were considered at risk for events.

Bottom line: Based on a conceptual model of 2004 hospitalized patients, eliminating preventable adverse events could have saved $16.6 billion, while eliminating redundant tests could have saved another $8 billion.

Citation: Jha AK, Chan DC, Ridgway AB, Franz C, Bates DW. Improving safety and eliminating redundant tests: cutting costs in U.S. hospitals. Health Aff (Millwood). 2009;28(5):1475-1484.

Trauma Patients with Pulmonary Embolism Might Not Have DVT on Imaging of Lower Extremities

Clinical question: What is the relationship between acute DVT and pulmonary embolism (PE) in trauma patients?

Background: Major trauma is associated with an increased risk of acute DVT and PE. It is assumed that the majority of PEs arise from DVTs in the lower extremities. Definitive evidence demonstrating that PEs form in situ rather than embolize from leg veins could impact indications for inferior vena cava filters.

Study design: Retrospective chart review.

Setting: Academic Level 1 trauma center in Boston.

Synopsis: The medical records of 247 trauma patients with suspected PE who underwent CT angiography of the lungs and simultaneous CT venography of the pelvis and lower extremities from January 2004 to December 2007 were reviewed. High-risk patients also underwent weekly screening with duplex ultrasonagraphy of the legs.

PE was diagnosed in 46 patients (19%) and DVT in 18 patients (7%). Anticoagulant prophylaxis had been administered to 96% and 78% of the patients with PE and DVT, respectively. PE was diagnosed a median of 5.5 days after admission (range 0-40 days) and the majority (61%) were in segmental or subsegmental branches, rather than in the main or lobar pulmonary arteries (39%). Only seven of the 46 patients (15%) diagnosed with PE also had a pelvic or lower-extremity DVT on simultaneous imaging with CT venography.

Bottom line: Trauma patients with PE often do not have a DVT at the time of diagnosis, though it remains unknown whether this is due to in-situ pulmonary thrombosis or complete embolization from the lower extremities.

Citation: Velmahos GC, Spaniolas K, Tabbara M, et al. Pulmonary embolism and deep venous thrombosis in trauma: are they related? Arch Surg. 2009;144:928-932.

Cancer Guideline for VTE Prophylaxis for Inpatients and Long-Term Treatment With Low-Molecular-Weight Heparin for Acute VTE

Clinical question: On what aspects of VTE management in cancer patients are there consensus among the major guideline panels?

Background: VTE is a common and serious complication of cancer. Patients might be hypercoagulable due to prothrombotic mediators released or mediated by tumor cells, chemotherapeutic agents, debility, central venous catheters, hospitalizations, or surgical procedures. The optimal management often is problematic due to uncertain benefit and risk of bleeding.

Study design: Review of major guideline statements.

Synopsis: The authors examined five VTE guidelines of American and European cancer societies. Each guideline was reviewed to determine the main recommendations and whether there was consensus on key aspects of anticoagulant management.

The study authors concluded that consensus was reached on most key recommendations:

• VTE prophylaxis in hospitalized medical patients. All five guidelines recommend the use of prophylaxis, though some guidelines recommend anticoagulant prophylaxis for all inpatients in the absence of contraindications and some recommend limiting prophylaxis to immobilized patients. All five recommend the use of either unfractionated heparin, low-molecular-weight heparin (LMWH), or fondaparinux.
• VTE prevention in cancer patients undergoing surgery. All five guidelines recommend anticoagulant prophylaxis in the absence of contraindications and extending prophylaxis approximately four weeks after major surgery.
• VTE prophylaxis in cancer patients with central venous catheters. Not recommended.
• VTE prophylaxis in ambulatory cancer patients without central venous catheters. Recommended only for multiple myeloma patients receiving a thalidomide-lenalidomide regimen.
• Long-term treatment of acute VTE in cancer patients. All five guidelines recommend initial treatment with LMWH for at least three to six months, followed by indefinite treatment with LMWH or a vitamin K antagonist.

Bottom line: Major guideline panels agree on key aspects of VTE management for cancer patients, including the use of prophylaxis for hospitalized medical and surgical patients and the use of long-term LMWH treatment for cancer patients with acute VTE.

Citation: Khorana AA, Streiff MB, Farge D, et al. Venous thromboembolism prophylaxis and treatment in cancer: a consensus statement of major guidelines panels and call to action. J Clin Oncol. 2009; 27(29):4919-4926.

Discontinuation of Beta Blockers Increases Risk of Postoperative Myocardial Infarction and Death

Clinical question: Does perioperative beta-blocker discontinuation affect postoperative myocardial infarction (MI) in low-risk patients undergoing joint arthroplasty?

Background: Recent trials show no benefit of perioperative beta blockers in reducing the incidence of perioperative myocardial infarctions (POMI) in low-risk patients. This retrospective study examined the impact of continuing or discontinuing beta blockers and the occurrence of POMI in patients undergoing elective joint arthroplasties.

Study design: Retrospective chart review.

Setting: Large academic center in Ottawa, Canada.

Synopsis: Medical records for 5,178 patients undergoing elective hip or knee arthroplasty from January 2002 to June 2006 were included in the review. The primary outcome was POMI, defined as an increased troponin level. Patients were divided into three groups: beta blocker prescribed on post-operative day (POD) zero and continued for one week or until discharge; beta blocker prescribed on POD zero and discontinued at any time in the first week; and no beta blocker on POD 0.

Beta blockers were continued in 992 patients and discontinued in 252 patients. The rate of POMI and death increased in the beta-blocker discontinuation group (odds ratio 2.0 [1.1-3.9] and 2.0 [1.1-3.9], respectively). This association persisted after adjustment for cardiac risk using a validated risk score.

The study was limited by the fact that the control group did not include patients who were on a beta blocker at home, thus potentially increasing the number of events in this group. The discontinuation beta blocker group had an increased baseline risk for POMI. The reason for discontinuing the beta blocker was not known, and cessation of beta blocker could have been due to an acute event.

Bottom line: This study adds support to the American College of Cardiology and American Heart Association (ACC/AHA) guidelines, which recommend continuation of beta-blocker therapy in the perioperative period.

Citation: Van Klei WA, Bryson GL, Yang H, Forster AJ. Effect of beta-blocker prescription on the incidence of postoperative myocardial infarction after hip and knee arthroplasty. Anesthesiology. 2009;111(4):717-724.

Lower Perioperative Mortality with Endovascular Vs. Open Abdominal Aortic Aneurysm Repair

Clinical question: How do perioperative and long-term morbidity and mortality compare in endovascular and open repair of abdominal aortic aneurysm (AAA)?

Background: Open AAA repair has relatively high perioperative mortality. Endovascular repair was developed as a less-invasive option and has been shown to reduce inpatient perioperative mortality, length of hospital stay, and ICU requirement. However, data suggest it leads to more frequent reinterventions and the same mortality rate as open repair at two years.

Study design: Randomized clinical trial.

Setting: Veterans Affairs medical centers.

Synopsis: The study randomized 881 veterans who planned to have elective AAA repair and were eligible for both endovascular and open repair. This is a planned, two-year interim report in a nine-year study.

Perioperative mortality was 0.5% in the endovascular repair group, compared with 3.0% in the open repair group. However, this difference in mortality was not statistically significant at two years. The endovascular repair group experienced shorter procedure and mechanical ventilation time, decreased hospital and ICU stay, and lower rate of blood transfusions.

Overall, there was no difference between the groups for major morbidity, procedure failure, need for secondary therapeutic intervention, quality of life, or erectile dysfunction. More data on long-term comparison of these two interventions will be available at the conclusion of this study.

Bottom line: Endovascular repair of AAA has lower perioperative mortality than open repair but did not lead to improved morbidity or mortality at two years.

Citation: Lederle FA, Freischlag JA, Kyriakides TC, et al. Outcomes following endovascular vs. open repair of abdominal aortic aneurysm: a randomized trial. JAMA. 2009;302 (14):1535-1542.

OTC Analgesics Not Associated with Acute Decompensation in Cirrhotic Patients

Clinical question: Do over-the-counter (OTC) analgesics lead to acute hepatic decompensation among patients with cirrhosis?

Background: In theory, intake of acetaminophen and/or nonsteroidal anti-inflammatory drugs (NSAIDs) can worsen hepatic function and lead to complications among cirrhotic patients. The role of OTC analgesics in potentially triggering acute hepatic decompensation among cirrhotic patients has not been studied.

Study design: Prospective case-control study.

Setting: Two tertiary-care hospitals.

Synopsis: Cirrhotic patients hospitalized for acute liver decompensation were compared with compensated cirrhotic patients in the liver clinic (cirrhotic controls) and with randomly selected, noncirrhotic patients who were simultaneously hospitalized (noncirrhotic controls). Data collected through questionnaires included quantity and dose of OTC analgesics used and alcohol consumption in the past 30 days.

Thirty-five percent of the hospitalized cirrhotic patients, 52% of the cirrhotic controls, and 70% of the noncirrhotic controls used OTC analgesics. At doses lower than those recommended, acetaminophen is not associated with acute liver decompensation among cirrhotic patients, even with recent alcohol use. However, NSAIDs taken by the cirrhotic patients, when compared to control subjects, were in larger doses and used for a longer duration, suggesting NSAIDs may have contributed to the acute decompensation.

Study limitations include the nature of the study design, reliance on the patient’s recall of OTC analgesic use, and obtaining other possible causes of decompensation, such as herbal supplement intake or compliance with diuretics or dietary indiscretion.

Bottom line: Acetaminophen at doses lower than recommended is not associated with adverse complications in cirrhotic patients, but NSAIDs are possibly associated with acute decompensation.

Citation: Khalid SK, Lane J, Navarro V, Garcia-Tsao G. Use of over-the-counter analgesics is not associated with acute decompensation in patients with cirrhosis. Clin Gastroenterol Hepatol. 2009;7(9):994-999.

Cardiovascular Disease and Risk of Hip Fracture

Clinical question: Is the diagnosis of cardiovascular disease (CVD) associated with the risk of subsequent hip fracture?

Background: Osteoporosis and CVD are regarded as independent, age-related conditions. However, recent research suggests that the bone and vascular systems share common regulatory mechanisms. Stroke is a known risk factor for hip fractures, and bisphosphonates have been shown to prevent atherosclerosis and reduce total mortality rate.

Study design: Cohort study.

Setting: Swedish National Patient Registry.

Synopsis: The study identified 31,936 Swedish twins born from 1914 to 1944. This cohort was followed up to age 50, and time-dependent exposures using Cox-proportional hazard regression models were evaluated.

Times to hip fracture after CVD diagnosis were isolated. Crude absolute rate of hip fractures (per 1,000 person-years) was 12.6 after diagnosis of heart failure, 12.6 after a stroke, 6.6 after peripheral atherosclerosis, and 5.2 after ischemic heart disease (IHD), compared with 1.2 per 1,000 person-years without a CVD diagnosis. Multivariable-adjusted hazard ratio (HR) of hip fracture after heart failure was 4.40 (95% CI, 3.43-5.63); after a stroke was 5.09 (95% CI, 4.18-6.20); after peripheral atherosclerosis was 3.20 (CI, 2.28-4.50); and after an IHD event was 2.32 (CI, 1.91-2.84).

Identical twins even without heart failure and stroke also had an increased risk of hip fracture if their twin had been diagnosed with these diseases.

Bottom line: Cardiovascular disease is significantly associated with risk of subsequent hip fracture, and genetic factors probably play a role in the association.

Citation: Sennerby U, Melhus H, Gedeborg R, et al. Cardiovascular diseases and risk of hip fracture. JAMA. 2009;302(15):1666-1673. TH

PEDIATRIC HM Literature

By Mark Shen, MD

Variation in the Treatment of Henoch-Schönlein Purpura

Reviewed by Pediatric Editor Mark Shen, MD, medical director of hospital medicine at Dell Children’s Medical Center, Austin, Texas.

Clinical question: What is the degree of variation in the inpatient management of Henoch-Schönlein purpura (HSP)?

Background: HSP is the most common pediatric vasculitis, but there are no consensus recommendations or guidelines for treatment. The amount of variation in the pharmacologic management of this disease is unknown.

Study design: Retrospective database analysis.

Setting: Thirty-six children’s hospitals affiliated with the Child Health Corporation of America.

Synopsis: The Pediatric Health Information (PHIS) database was sampled for children younger than 18 years of age with an ICD-9-CM code of HSP and discharge from a hospital that submitted appropriate data from 2000 to 2007. Only index admissions were included, and children with coexisting rheumatic conditions were excluded, for a total of 1,988 subjects.

Logistic regression analysis was used to examine the effects of patient-level standardization on hospital-level rates of therapy and the degree to which variation across hospitals occurred beyond what would be expected after standardization.

Hospital-level variation in medication use was significant (P<0.001) for corticosteroids, opiates, and nonsteroidal anti-inflammatory drugs (NSAIDs), even after adjustment for severity and age at presentation.

Although variation in management is not surprising, the significant degree to which this occurred at the hospital level suggests that local institutional culture plays a dominant role in decision-making. The use of the PHIS database allows for analysis of a large population that would be otherwise difficult to study. However, significant numbers of HSP patients do not require hospitalization, and the study results might substantially over- or underestimate practice patterns. Collaborative efforts to better define optimal management of HSP are needed.

Bottom line: A significant degree of hospital-level variation exists in the inpatient management of HSP.

Citation: Weiss PF, Klink AJ, Hexem K, et al. Variation in inpatient therapy and diagnostic evaluation of children with henoch schönlein purpura. J Pediatr. 2009;155(6):812-818.e1.

In This Edition

Literature at a Glance

A guide to this month’s studies

• Perioperative oxygen use and infection rate.
• Effects of adverse events and healthcare costs
• Impact of DVT on PE rates in trauma
• VTE prevention and treatment in cancer patients
• Effect of perioperative beta-blocker discontinuation
• Endovascular vs. open AAA repair
• OTC analgesics in patients with hepatic dysfunction
• Cardiovascular disease and risk of hip fracture

High Perioperative Oxygen Fraction Does Not Improve Surgical-Site Infection Frequency after Abdominal Surgery

Clinical question: Does the use of 80% oxygen perioperatively in abdominal surgery decrease the frequency of surgical-site infection within 14 days without increasing the rate of pulmonary complications?

Background: Low oxygen tension in wounds can negatively impact immune response and healing. Increasing inspiratory oxygen fraction during the perioperative period translates into higher wound oxygen tension. However, the benefit of increased oxygen fraction therapy in abdominal surgery healing and complications is not clear, nor is the frequency of pulmonary complications.

Study design: Patient- and observer-blinded clinical trial.

Setting: Fourteen Danish hospitals from October 2006 to October 2008.

Synopsis: Patients were randomized to receive a fraction of inspired oxygen (FIO₂) of 0.80 or 0.30. The primary outcome—surgical-site infection in the superficial or deep wound or intra-abdominal cavity within 14 days of surgery—was defined using Centers for Disease Control and Prevention (CDC) criteria. Secondary outcomes included pulmonary complications within 14 days (pneumonia, atelectasis, or respiratory failure), 30-day mortality, duration of post-op course, ICU stay within 14 days post-op, and any abdominal operation within 14 days. The 1,386 patients were enrolled in the intention-to-treat analysis.

Infection occurred in 19.1% of patients given 0.80 FIO2 and in 20.1% of patients given 0.30 FIO2; odds ratio of 0.94 (95% CI 0.72 to 1.22; P=0.64). Numbers of pulmonary complications were not significantly different between the groups.

This trial included acute and nonacute laparotomies with followup for adverse outcomes. Study limitations included the inability to ensure that both groups received timely antibiotics and prevention for hypothermia. Of patients in the 30% FIO2 group, 7.3% required higher oxygen administration. Additionally, infection might have been underestimated in 11.3% of patients who were not followed up on between days 13 and 30.

Bottom line: High oxygen concentration administered during and after laparotomy did not lead to fewer surgical site infections, nor did it significantly increase the frequency of pulmonary complications or death.

Citation: Meyhoff CS, Wetterslev J, Jorgensen LN, et al. Effect of high perioperative oxygen fraction on surgical site infection and pulmonary complications after abdominal surgery: the PROXI randomized clinical trial. JAMA. 2009;302(14):1543-1550.

Eliminating Adverse Events and Redundant Tests Could Generate U.S. Healthcare Savings

Clinical question: Using available data, what is the estimated cost savings of eliminating adverse events and avoiding redundant tests?

Background: Reimbursement schemes are changing such that hospitals are reimbursed less for some adverse events. This financial disincentive is expected to spark interest in improved patient safety. The authors sought to model the cost savings generated by eliminating redundant testing and adverse events from literature-based estimates.

Study design: Development of conceptual model to identify common or costly adverse events, redundant tests, and simulated costs.

Setting: Literature review, expert opinion, data from safety organizations and epidemiologic studies, and patient data from the 2004 National Inpatient Data Sample.

Synopsis: The conceptual model identified 5.7 million adverse events in U.S. hospitals, of which 3 million were considered preventable. The most common events included hospital-acquired infections (82% preventable), adverse drug events (26%), falls (33%), and iatrogenic thromboembolic events (62%). The calculated cost savings totaled $16.6 billion (5.5% of total inpatient costs) for adverse events and $8.2 billion for the elimination of redundant tests. When looking at hospital subtypes, the greatest savings would come from major teaching hospitals.

This study is limited by its use of published and heterogeneous data spanning a 15-year period. The authors did not include events for which there was no epidemiologic or cost data. As hospital-care changes and technology is adopted, it is uncertain how this changes the costs, prevalence, and the preventable nature of these events. The model was not consistently able to identifying high- and low-risk patients. For instance, in some models, all patients were considered at risk for events.

Bottom line: Based on a conceptual model of 2004 hospitalized patients, eliminating preventable adverse events could have saved $16.6 billion, while eliminating redundant tests could have saved another $8 billion.

Citation: Jha AK, Chan DC, Ridgway AB, Franz C, Bates DW. Improving safety and eliminating redundant tests: cutting costs in U.S. hospitals. Health Aff (Millwood). 2009;28(5):1475-1484.

Trauma Patients with Pulmonary Embolism Might Not Have DVT on Imaging of Lower Extremities

Clinical question: What is the relationship between acute DVT and pulmonary embolism (PE) in trauma patients?

Background: Major trauma is associated with an increased risk of acute DVT and PE. It is assumed that the majority of PEs arise from DVTs in the lower extremities. Definitive evidence demonstrating that PEs form in situ rather than embolize from leg veins could impact indications for inferior vena cava filters.

Study design: Retrospective chart review.

Setting: Academic Level 1 trauma center in Boston.

Synopsis: The medical records of 247 trauma patients with suspected PE who underwent CT angiography of the lungs and simultaneous CT venography of the pelvis and lower extremities from January 2004 to December 2007 were reviewed. High-risk patients also underwent weekly screening with duplex ultrasonagraphy of the legs.

PE was diagnosed in 46 patients (19%) and DVT in 18 patients (7%). Anticoagulant prophylaxis had been administered to 96% and 78% of the patients with PE and DVT, respectively. PE was diagnosed a median of 5.5 days after admission (range 0-40 days) and the majority (61%) were in segmental or subsegmental branches, rather than in the main or lobar pulmonary arteries (39%). Only seven of the 46 patients (15%) diagnosed with PE also had a pelvic or lower-extremity DVT on simultaneous imaging with CT venography.

Bottom line: Trauma patients with PE often do not have a DVT at the time of diagnosis, though it remains unknown whether this is due to in-situ pulmonary thrombosis or complete embolization from the lower extremities.

Citation: Velmahos GC, Spaniolas K, Tabbara M, et al. Pulmonary embolism and deep venous thrombosis in trauma: are they related? Arch Surg. 2009;144:928-932.

Cancer Guideline for VTE Prophylaxis for Inpatients and Long-Term Treatment With Low-Molecular-Weight Heparin for Acute VTE

Clinical question: On what aspects of VTE management in cancer patients are there consensus among the major guideline panels?

Background: VTE is a common and serious complication of cancer. Patients might be hypercoagulable due to prothrombotic mediators released or mediated by tumor cells, chemotherapeutic agents, debility, central venous catheters, hospitalizations, or surgical procedures. The optimal management often is problematic due to uncertain benefit and risk of bleeding.

Study design: Review of major guideline statements.

Synopsis: The authors examined five VTE guidelines of American and European cancer societies. Each guideline was reviewed to determine the main recommendations and whether there was consensus on key aspects of anticoagulant management.

The study authors concluded that consensus was reached on most key recommendations:

• VTE prophylaxis in hospitalized medical patients. All five guidelines recommend the use of prophylaxis, though some guidelines recommend anticoagulant prophylaxis for all inpatients in the absence of contraindications and some recommend limiting prophylaxis to immobilized patients. All five recommend the use of either unfractionated heparin, low-molecular-weight heparin (LMWH), or fondaparinux.
• VTE prevention in cancer patients undergoing surgery. All five guidelines recommend anticoagulant prophylaxis in the absence of contraindications and extending prophylaxis approximately four weeks after major surgery.
• VTE prophylaxis in cancer patients with central venous catheters. Not recommended.
• VTE prophylaxis in ambulatory cancer patients without central venous catheters. Recommended only for multiple myeloma patients receiving a thalidomide-lenalidomide regimen.
• Long-term treatment of acute VTE in cancer patients. All five guidelines recommend initial treatment with LMWH for at least three to six months, followed by indefinite treatment with LMWH or a vitamin K antagonist.

Bottom line: Major guideline panels agree on key aspects of VTE management for cancer patients, including the use of prophylaxis for hospitalized medical and surgical patients and the use of long-term LMWH treatment for cancer patients with acute VTE.

Citation: Khorana AA, Streiff MB, Farge D, et al. Venous thromboembolism prophylaxis and treatment in cancer: a consensus statement of major guidelines panels and call to action. J Clin Oncol. 2009; 27(29):4919-4926.

Discontinuation of Beta Blockers Increases Risk of Postoperative Myocardial Infarction and Death

Clinical question: Does perioperative beta-blocker discontinuation affect postoperative myocardial infarction (MI) in low-risk patients undergoing joint arthroplasty?

Background: Recent trials show no benefit of perioperative beta blockers in reducing the incidence of perioperative myocardial infarctions (POMI) in low-risk patients. This retrospective study examined the impact of continuing or discontinuing beta blockers and the occurrence of POMI in patients undergoing elective joint arthroplasties.

Study design: Retrospective chart review.

Setting: Large academic center in Ottawa, Canada.

Synopsis: Medical records for 5,178 patients undergoing elective hip or knee arthroplasty from January 2002 to June 2006 were included in the review. The primary outcome was POMI, defined as an increased troponin level. Patients were divided into three groups: beta blocker prescribed on post-operative day (POD) zero and continued for one week or until discharge; beta blocker prescribed on POD zero and discontinued at any time in the first week; and no beta blocker on POD 0.

Beta blockers were continued in 992 patients and discontinued in 252 patients. The rate of POMI and death increased in the beta-blocker discontinuation group (odds ratio 2.0 [1.1-3.9] and 2.0 [1.1-3.9], respectively). This association persisted after adjustment for cardiac risk using a validated risk score.

The study was limited by the fact that the control group did not include patients who were on a beta blocker at home, thus potentially increasing the number of events in this group. The discontinuation beta blocker group had an increased baseline risk for POMI. The reason for discontinuing the beta blocker was not known, and cessation of beta blocker could have been due to an acute event.

Bottom line: This study adds support to the American College of Cardiology and American Heart Association (ACC/AHA) guidelines, which recommend continuation of beta-blocker therapy in the perioperative period.

Citation: Van Klei WA, Bryson GL, Yang H, Forster AJ. Effect of beta-blocker prescription on the incidence of postoperative myocardial infarction after hip and knee arthroplasty. Anesthesiology. 2009;111(4):717-724.

Lower Perioperative Mortality with Endovascular Vs. Open Abdominal Aortic Aneurysm Repair

Clinical question: How do perioperative and long-term morbidity and mortality compare in endovascular and open repair of abdominal aortic aneurysm (AAA)?

Background: Open AAA repair has relatively high perioperative mortality. Endovascular repair was developed as a less-invasive option and has been shown to reduce inpatient perioperative mortality, length of hospital stay, and ICU requirement. However, data suggest it leads to more frequent reinterventions and the same mortality rate as open repair at two years.

Study design: Randomized clinical trial.

Setting: Veterans Affairs medical centers.

Synopsis: The study randomized 881 veterans who planned to have elective AAA repair and were eligible for both endovascular and open repair. This is a planned, two-year interim report in a nine-year study.

Perioperative mortality was 0.5% in the endovascular repair group, compared with 3.0% in the open repair group. However, this difference in mortality was not statistically significant at two years. The endovascular repair group experienced shorter procedure and mechanical ventilation time, decreased hospital and ICU stay, and lower rate of blood transfusions.

Overall, there was no difference between the groups for major morbidity, procedure failure, need for secondary therapeutic intervention, quality of life, or erectile dysfunction. More data on long-term comparison of these two interventions will be available at the conclusion of this study.

Bottom line: Endovascular repair of AAA has lower perioperative mortality than open repair but did not lead to improved morbidity or mortality at two years.

Citation: Lederle FA, Freischlag JA, Kyriakides TC, et al. Outcomes following endovascular vs. open repair of abdominal aortic aneurysm: a randomized trial. JAMA. 2009;302 (14):1535-1542.

OTC Analgesics Not Associated with Acute Decompensation in Cirrhotic Patients

Clinical question: Do over-the-counter (OTC) analgesics lead to acute hepatic decompensation among patients with cirrhosis?

Background: In theory, intake of acetaminophen and/or nonsteroidal anti-inflammatory drugs (NSAIDs) can worsen hepatic function and lead to complications among cirrhotic patients. The role of OTC analgesics in potentially triggering acute hepatic decompensation among cirrhotic patients has not been studied.

Study design: Prospective case-control study.

Setting: Two tertiary-care hospitals.

Synopsis: Cirrhotic patients hospitalized for acute liver decompensation were compared with compensated cirrhotic patients in the liver clinic (cirrhotic controls) and with randomly selected, noncirrhotic patients who were simultaneously hospitalized (noncirrhotic controls). Data collected through questionnaires included quantity and dose of OTC analgesics used and alcohol consumption in the past 30 days.

Thirty-five percent of the hospitalized cirrhotic patients, 52% of the cirrhotic controls, and 70% of the noncirrhotic controls used OTC analgesics. At doses lower than those recommended, acetaminophen is not associated with acute liver decompensation among cirrhotic patients, even with recent alcohol use. However, NSAIDs taken by the cirrhotic patients, when compared to control subjects, were in larger doses and used for a longer duration, suggesting NSAIDs may have contributed to the acute decompensation.

Study limitations include the nature of the study design, reliance on the patient’s recall of OTC analgesic use, and obtaining other possible causes of decompensation, such as herbal supplement intake or compliance with diuretics or dietary indiscretion.

Bottom line: Acetaminophen at doses lower than recommended is not associated with adverse complications in cirrhotic patients, but NSAIDs are possibly associated with acute decompensation.

Citation: Khalid SK, Lane J, Navarro V, Garcia-Tsao G. Use of over-the-counter analgesics is not associated with acute decompensation in patients with cirrhosis. Clin Gastroenterol Hepatol. 2009;7(9):994-999.

Cardiovascular Disease and Risk of Hip Fracture

Clinical question: Is the diagnosis of cardiovascular disease (CVD) associated with the risk of subsequent hip fracture?

Background: Osteoporosis and CVD are regarded as independent, age-related conditions. However, recent research suggests that the bone and vascular systems share common regulatory mechanisms. Stroke is a known risk factor for hip fractures, and bisphosphonates have been shown to prevent atherosclerosis and reduce total mortality rate.

Study design: Cohort study.

Setting: Swedish National Patient Registry.

Synopsis: The study identified 31,936 Swedish twins born from 1914 to 1944. This cohort was followed up to age 50, and time-dependent exposures using Cox-proportional hazard regression models were evaluated.

Times to hip fracture after CVD diagnosis were isolated. Crude absolute rate of hip fractures (per 1,000 person-years) was 12.6 after diagnosis of heart failure, 12.6 after a stroke, 6.6 after peripheral atherosclerosis, and 5.2 after ischemic heart disease (IHD), compared with 1.2 per 1,000 person-years without a CVD diagnosis. Multivariable-adjusted hazard ratio (HR) of hip fracture after heart failure was 4.40 (95% CI, 3.43-5.63); after a stroke was 5.09 (95% CI, 4.18-6.20); after peripheral atherosclerosis was 3.20 (CI, 2.28-4.50); and after an IHD event was 2.32 (CI, 1.91-2.84).

Identical twins even without heart failure and stroke also had an increased risk of hip fracture if their twin had been diagnosed with these diseases.

Bottom line: Cardiovascular disease is significantly associated with risk of subsequent hip fracture, and genetic factors probably play a role in the association.

Citation: Sennerby U, Melhus H, Gedeborg R, et al. Cardiovascular diseases and risk of hip fracture. JAMA. 2009;302(15):1666-1673. TH

PEDIATRIC HM Literature

By Mark Shen, MD

Variation in the Treatment of Henoch-Schönlein Purpura

Reviewed by Pediatric Editor Mark Shen, MD, medical director of hospital medicine at Dell Children’s Medical Center, Austin, Texas.

Clinical question: What is the degree of variation in the inpatient management of Henoch-Schönlein purpura (HSP)?

Background: HSP is the most common pediatric vasculitis, but there are no consensus recommendations or guidelines for treatment. The amount of variation in the pharmacologic management of this disease is unknown.

Study design: Retrospective database analysis.

Setting: Thirty-six children’s hospitals affiliated with the Child Health Corporation of America.

Synopsis: The Pediatric Health Information (PHIS) database was sampled for children younger than 18 years of age with an ICD-9-CM code of HSP and discharge from a hospital that submitted appropriate data from 2000 to 2007. Only index admissions were included, and children with coexisting rheumatic conditions were excluded, for a total of 1,988 subjects.

Logistic regression analysis was used to examine the effects of patient-level standardization on hospital-level rates of therapy and the degree to which variation across hospitals occurred beyond what would be expected after standardization.

Hospital-level variation in medication use was significant (P<0.001) for corticosteroids, opiates, and nonsteroidal anti-inflammatory drugs (NSAIDs), even after adjustment for severity and age at presentation.

Although variation in management is not surprising, the significant degree to which this occurred at the hospital level suggests that local institutional culture plays a dominant role in decision-making. The use of the PHIS database allows for analysis of a large population that would be otherwise difficult to study. However, significant numbers of HSP patients do not require hospitalization, and the study results might substantially over- or underestimate practice patterns. Collaborative efforts to better define optimal management of HSP are needed.

Bottom line: A significant degree of hospital-level variation exists in the inpatient management of HSP.

Citation: Weiss PF, Klink AJ, Hexem K, et al. Variation in inpatient therapy and diagnostic evaluation of children with henoch schönlein purpura. J Pediatr. 2009;155(6):812-818.e1.