Q&A

CLINICAL QUESTION: In a low-risk primary care population, is warfarin more effective than aspirin in preventing thromboembolism?

BACKGROUND: Studies of patients with atrial fibrillation have shown that warfarin is more effective than aspirin in reducing the risk of thromboembolism. It is unclear whether this finding is generalizable to primary care settings, since patients in these particular studies may be at higher risk than the typical family practice patient. The purpose of this study was to compare the rates of thromboembolism and the side effects of aspirin and warfarin in low-risk primary care patients with atrial fibrillation.

POPULATION STUDIED: A total of 729 patients aged 60 years and older were recruited from 284 general practices in the Netherlands. The pulse of all older patients visiting these practices was taken to identify those with atrial fibrillation for this study. Atrial fibrillation was confirmed by electrocardiogram. Patients were excluded if they had rheumatic valvular disease; previous stroke or systemic embolism; treatable causes of atrial fibrillation; contraindications to anticoagulation; or a history of recent warfarin use, myocardial surgery, or infarction.

STUDY DESIGN AND VALIDITY: This was a randomized single-blind controlled trial. Patients eligible for standard intensity warfarin were randomly assigned to either regular-dose warfarin (to achieve an international normalized ratio [INR] of 2.5 - 3.5), low-intensity warfarin (INR = 1.1 - 1.6), or aspirin 150 mg per day (stratum 1). Patients ineligible for standard-dose anticoagulation were randomly assigned to low-dose warfarin or aspirin (stratum 2). Patients were unaware of their intensity of anticoagulant. However, they were not blinded as to which medication (aspirin or warfarin) they were taking. Evaluators were unaware of treatment. This study enrolled patients at low baseline risk of embolism. Twenty-eight percent of the patients were taking aspirin before entering the trial. Although aspirin use was stopped before randomization, this group may have been at lower risk for thromboembolism. It is likely that this study had a predominantly white population. The results may not be generalizable to blacks, who have a higher rate of strokes than whites for any given level of blood pressure. Sample size calculations showed that the study had sufficient statistical power to compare low-intensity anticoagulation (INR = 1.1 - 1.6)—but not standard anticoagulation—to aspirin. The low rate of events in the group as a whole (5.5% per year) made it difficult to find differences in outcome between the treatment groups. In addition, the study was too small to detect a difference in outcome between the standard anticoagulation and aspirin groups.

OUTCOMES MEASURED: The primary outcomes were stroke, systemic embolism, major hemorrhage, and vascular death. Secondary outcomes were nonfatal myocardial infarction, retinal infarction, transient ischemic attack, minor bleeding, and nonvascular death. results n During an average follow-up period of 2.7 years, 108 patients (5.5% per year) suffered a major event. This rate of stroke and other events is lower than in the typical populations studied in the past. There was no difference in the incidence of these major events in patients receiving aspirin or low or standard anticoagulation. Nonvascular death was less common in the low-anticoagulation group than in the aspirin group (hazard ratio = 0.41; 95% confidence interval, 0.20 - 0.82). Major bleeding events occurred in 0.7% of patients per year; this rate was not different among the 3 groups.

CLINICAL QUESTION: In a low-risk primary care population, is warfarin more effective than aspirin in preventing thromboembolism?

BACKGROUND: Studies of patients with atrial fibrillation have shown that warfarin is more effective than aspirin in reducing the risk of thromboembolism. It is unclear whether this finding is generalizable to primary care settings, since patients in these particular studies may be at higher risk than the typical family practice patient. The purpose of this study was to compare the rates of thromboembolism and the side effects of aspirin and warfarin in low-risk primary care patients with atrial fibrillation.

POPULATION STUDIED: A total of 729 patients aged 60 years and older were recruited from 284 general practices in the Netherlands. The pulse of all older patients visiting these practices was taken to identify those with atrial fibrillation for this study. Atrial fibrillation was confirmed by electrocardiogram. Patients were excluded if they had rheumatic valvular disease; previous stroke or systemic embolism; treatable causes of atrial fibrillation; contraindications to anticoagulation; or a history of recent warfarin use, myocardial surgery, or infarction.

STUDY DESIGN AND VALIDITY: This was a randomized single-blind controlled trial. Patients eligible for standard intensity warfarin were randomly assigned to either regular-dose warfarin (to achieve an international normalized ratio [INR] of 2.5 - 3.5), low-intensity warfarin (INR = 1.1 - 1.6), or aspirin 150 mg per day (stratum 1). Patients ineligible for standard-dose anticoagulation were randomly assigned to low-dose warfarin or aspirin (stratum 2). Patients were unaware of their intensity of anticoagulant. However, they were not blinded as to which medication (aspirin or warfarin) they were taking. Evaluators were unaware of treatment. This study enrolled patients at low baseline risk of embolism. Twenty-eight percent of the patients were taking aspirin before entering the trial. Although aspirin use was stopped before randomization, this group may have been at lower risk for thromboembolism. It is likely that this study had a predominantly white population. The results may not be generalizable to blacks, who have a higher rate of strokes than whites for any given level of blood pressure. Sample size calculations showed that the study had sufficient statistical power to compare low-intensity anticoagulation (INR = 1.1 - 1.6)—but not standard anticoagulation—to aspirin. The low rate of events in the group as a whole (5.5% per year) made it difficult to find differences in outcome between the treatment groups. In addition, the study was too small to detect a difference in outcome between the standard anticoagulation and aspirin groups.

OUTCOMES MEASURED: The primary outcomes were stroke, systemic embolism, major hemorrhage, and vascular death. Secondary outcomes were nonfatal myocardial infarction, retinal infarction, transient ischemic attack, minor bleeding, and nonvascular death. results n During an average follow-up period of 2.7 years, 108 patients (5.5% per year) suffered a major event. This rate of stroke and other events is lower than in the typical populations studied in the past. There was no difference in the incidence of these major events in patients receiving aspirin or low or standard anticoagulation. Nonvascular death was less common in the low-anticoagulation group than in the aspirin group (hazard ratio = 0.41; 95% confidence interval, 0.20 - 0.82). Major bleeding events occurred in 0.7% of patients per year; this rate was not different among the 3 groups.

CLINICAL QUESTION: In a low-risk primary care population, is warfarin more effective than aspirin in preventing thromboembolism?

BACKGROUND: Studies of patients with atrial fibrillation have shown that warfarin is more effective than aspirin in reducing the risk of thromboembolism. It is unclear whether this finding is generalizable to primary care settings, since patients in these particular studies may be at higher risk than the typical family practice patient. The purpose of this study was to compare the rates of thromboembolism and the side effects of aspirin and warfarin in low-risk primary care patients with atrial fibrillation.

POPULATION STUDIED: A total of 729 patients aged 60 years and older were recruited from 284 general practices in the Netherlands. The pulse of all older patients visiting these practices was taken to identify those with atrial fibrillation for this study. Atrial fibrillation was confirmed by electrocardiogram. Patients were excluded if they had rheumatic valvular disease; previous stroke or systemic embolism; treatable causes of atrial fibrillation; contraindications to anticoagulation; or a history of recent warfarin use, myocardial surgery, or infarction.

STUDY DESIGN AND VALIDITY: This was a randomized single-blind controlled trial. Patients eligible for standard intensity warfarin were randomly assigned to either regular-dose warfarin (to achieve an international normalized ratio [INR] of 2.5 - 3.5), low-intensity warfarin (INR = 1.1 - 1.6), or aspirin 150 mg per day (stratum 1). Patients ineligible for standard-dose anticoagulation were randomly assigned to low-dose warfarin or aspirin (stratum 2). Patients were unaware of their intensity of anticoagulant. However, they were not blinded as to which medication (aspirin or warfarin) they were taking. Evaluators were unaware of treatment. This study enrolled patients at low baseline risk of embolism. Twenty-eight percent of the patients were taking aspirin before entering the trial. Although aspirin use was stopped before randomization, this group may have been at lower risk for thromboembolism. It is likely that this study had a predominantly white population. The results may not be generalizable to blacks, who have a higher rate of strokes than whites for any given level of blood pressure. Sample size calculations showed that the study had sufficient statistical power to compare low-intensity anticoagulation (INR = 1.1 - 1.6)—but not standard anticoagulation—to aspirin. The low rate of events in the group as a whole (5.5% per year) made it difficult to find differences in outcome between the treatment groups. In addition, the study was too small to detect a difference in outcome between the standard anticoagulation and aspirin groups.

OUTCOMES MEASURED: The primary outcomes were stroke, systemic embolism, major hemorrhage, and vascular death. Secondary outcomes were nonfatal myocardial infarction, retinal infarction, transient ischemic attack, minor bleeding, and nonvascular death. results n During an average follow-up period of 2.7 years, 108 patients (5.5% per year) suffered a major event. This rate of stroke and other events is lower than in the typical populations studied in the past. There was no difference in the incidence of these major events in patients receiving aspirin or low or standard anticoagulation. Nonvascular death was less common in the low-anticoagulation group than in the aspirin group (hazard ratio = 0.41; 95% confidence interval, 0.20 - 0.82). Major bleeding events occurred in 0.7% of patients per year; this rate was not different among the 3 groups.

CLINICAL QUESTION: Which types of homemade spacers are most effective for inhaled b-agonist delivery in children with acute asthma exacerbations?

BACKGROUND: Multiple studies have demonstrated that for acute asthma exacerbations, metered dose inhalers (MDIs) with attached valved spacers are as effective, if not more effective, than nebulizers. Homemade spacers made from more readily available plastic bottles or polystyrene cups are sometimes substituted for conventional spacers. This study compares clinical outcomes using conventional spacers with 3 types of homemade spacers in children with acute asthma exacerbations.

POPULATION STUDIED: A total of 88 children aged 5 to 13 years who were experiencing an acute asthma attack were recruited from a Cape Town, South Africa, hospital emergency department. All subjects had a baseline peak expiratory flow rate (PEFR) between 20% and 80% predicted. The children were stratified by severity, with a PEFR of higher than 59% considered mild and a rate lower than 60% considered moderate. Patients who had used b-agonist medication within 4 hours of presentation were excluded.

STUDY DESIGN AND VALIDITY: This was a randomized trial comparing 4 spacer options: conventional valved spacers, plastic drink bottles sealed to the MDI with glue, plastic drink bottles that were not sealed, and polystyrene cups. To make the spacers, a heated wire in the shape of the MDI mouthpiece was pressed to the base of the bottle or cup to make a hole into which the MDI was inserted. Pulmonary function tests (PFTs) were performed at baseline and again after treatment with a b-agonist delivered by MDI and the assigned spacer type. Those patients who did not improve to at least 70% of predicted PEFR were given a nebulizer treatment with the same medication. Those nonresponders had a third set of PFTs after the nebulizer treatment. The investigators who evaluated outcomes were blinded to the treatment arm.

OUTCOMES MEASURED: The authors list 4 primary outcomes: change in a clinical asthma severity score, percent change in PFTs, failure to increase PEFR to higher than 70% of predicted, and change in PEFR with nebulizer treatment in nonresponders.

RESULTS: The pooled analysis, combining the mild and moderate groups, showed significant differences by spacer type for change in PFTs. For this combined group, the polystyrene cup resulted in minimal improvement in PFTs, while the other 3 spacer types showed significantly larger increases. When the mild severity group was analyzed separately, there were no significant differences by spacer type in any of the outcomes. For the moderate severity group, significant differences were found in change in PFTs, number of responders, and response to nebulizer for initial nonresponders. In all significant results in this moderate group, the conventional spacer performed the best and the polystyrene cup the worst, with the unsealed bottle somewhat more effective than the cup but less effective than the sealed bottle.

CLINICAL QUESTION: Which types of homemade spacers are most effective for inhaled b-agonist delivery in children with acute asthma exacerbations?

BACKGROUND: Multiple studies have demonstrated that for acute asthma exacerbations, metered dose inhalers (MDIs) with attached valved spacers are as effective, if not more effective, than nebulizers. Homemade spacers made from more readily available plastic bottles or polystyrene cups are sometimes substituted for conventional spacers. This study compares clinical outcomes using conventional spacers with 3 types of homemade spacers in children with acute asthma exacerbations.

POPULATION STUDIED: A total of 88 children aged 5 to 13 years who were experiencing an acute asthma attack were recruited from a Cape Town, South Africa, hospital emergency department. All subjects had a baseline peak expiratory flow rate (PEFR) between 20% and 80% predicted. The children were stratified by severity, with a PEFR of higher than 59% considered mild and a rate lower than 60% considered moderate. Patients who had used b-agonist medication within 4 hours of presentation were excluded.

STUDY DESIGN AND VALIDITY: This was a randomized trial comparing 4 spacer options: conventional valved spacers, plastic drink bottles sealed to the MDI with glue, plastic drink bottles that were not sealed, and polystyrene cups. To make the spacers, a heated wire in the shape of the MDI mouthpiece was pressed to the base of the bottle or cup to make a hole into which the MDI was inserted. Pulmonary function tests (PFTs) were performed at baseline and again after treatment with a b-agonist delivered by MDI and the assigned spacer type. Those patients who did not improve to at least 70% of predicted PEFR were given a nebulizer treatment with the same medication. Those nonresponders had a third set of PFTs after the nebulizer treatment. The investigators who evaluated outcomes were blinded to the treatment arm.

OUTCOMES MEASURED: The authors list 4 primary outcomes: change in a clinical asthma severity score, percent change in PFTs, failure to increase PEFR to higher than 70% of predicted, and change in PEFR with nebulizer treatment in nonresponders.

RESULTS: The pooled analysis, combining the mild and moderate groups, showed significant differences by spacer type for change in PFTs. For this combined group, the polystyrene cup resulted in minimal improvement in PFTs, while the other 3 spacer types showed significantly larger increases. When the mild severity group was analyzed separately, there were no significant differences by spacer type in any of the outcomes. For the moderate severity group, significant differences were found in change in PFTs, number of responders, and response to nebulizer for initial nonresponders. In all significant results in this moderate group, the conventional spacer performed the best and the polystyrene cup the worst, with the unsealed bottle somewhat more effective than the cup but less effective than the sealed bottle.

CLINICAL QUESTION: Which types of homemade spacers are most effective for inhaled b-agonist delivery in children with acute asthma exacerbations?

BACKGROUND: Multiple studies have demonstrated that for acute asthma exacerbations, metered dose inhalers (MDIs) with attached valved spacers are as effective, if not more effective, than nebulizers. Homemade spacers made from more readily available plastic bottles or polystyrene cups are sometimes substituted for conventional spacers. This study compares clinical outcomes using conventional spacers with 3 types of homemade spacers in children with acute asthma exacerbations.

POPULATION STUDIED: A total of 88 children aged 5 to 13 years who were experiencing an acute asthma attack were recruited from a Cape Town, South Africa, hospital emergency department. All subjects had a baseline peak expiratory flow rate (PEFR) between 20% and 80% predicted. The children were stratified by severity, with a PEFR of higher than 59% considered mild and a rate lower than 60% considered moderate. Patients who had used b-agonist medication within 4 hours of presentation were excluded.

STUDY DESIGN AND VALIDITY: This was a randomized trial comparing 4 spacer options: conventional valved spacers, plastic drink bottles sealed to the MDI with glue, plastic drink bottles that were not sealed, and polystyrene cups. To make the spacers, a heated wire in the shape of the MDI mouthpiece was pressed to the base of the bottle or cup to make a hole into which the MDI was inserted. Pulmonary function tests (PFTs) were performed at baseline and again after treatment with a b-agonist delivered by MDI and the assigned spacer type. Those patients who did not improve to at least 70% of predicted PEFR were given a nebulizer treatment with the same medication. Those nonresponders had a third set of PFTs after the nebulizer treatment. The investigators who evaluated outcomes were blinded to the treatment arm.

OUTCOMES MEASURED: The authors list 4 primary outcomes: change in a clinical asthma severity score, percent change in PFTs, failure to increase PEFR to higher than 70% of predicted, and change in PEFR with nebulizer treatment in nonresponders.

RESULTS: The pooled analysis, combining the mild and moderate groups, showed significant differences by spacer type for change in PFTs. For this combined group, the polystyrene cup resulted in minimal improvement in PFTs, while the other 3 spacer types showed significantly larger increases. When the mild severity group was analyzed separately, there were no significant differences by spacer type in any of the outcomes. For the moderate severity group, significant differences were found in change in PFTs, number of responders, and response to nebulizer for initial nonresponders. In all significant results in this moderate group, the conventional spacer performed the best and the polystyrene cup the worst, with the unsealed bottle somewhat more effective than the cup but less effective than the sealed bottle.

CLINICAL QUESTION: What are the concerns of women considering hormone replacement therapy (HRT)?

BACKGROUND: Counseling women about HRT is common in family practice, but available research has focused on risks and benefits from the perspective of providers. This qualitative study explores the perspective of patients.

POPULATION STUDIED: Twenty-six women who received new HRT prescriptions at a staff-model health maintenance organization were interviewed. Originally, 176 were identified as potential participants; reasons for exclusion included premenopausal state, previous use of estrogens, and use of topical estrogens. Thirty-four women refused to participate, 20 were never contacted, and 34 consented but were never interviewed because study data were sufficient. The median age was 53, 85% were white, and median annual income was $46,000. Fifty-four percent stated they had initiated discussion of HRT with their provider, and 81% had filled the HRT prescription at the time of interview.

STUDY DESIGN AND VALIDITY: This was a qualitative study. Each woman was interviewed for 1 hour, and the audiotape was transcribed and analyzed. Three judges used a consensus process to identify the variety of domains (specific concerns) the women mentioned. Another investigator assigned patient comments to specific domains. Interviewing new subjects stopped when no new information was being added.

OUTCOMES MEASURED: This qualitative study identified womens’ concerns about HRT. By design, a qualitative study can neither quantify the prevalence or strength of particular concerns nor address clinical outcomes.

RESULTS: Influences on the decision to begin HRT were broad and included provider opinion (96% of interviewees reported), media reports (81%), experiences and opinions of friends (77%), and experiences and opinions of family (65%). Specific clinical concerns cited included risk for breast cancer, having to take medication, prevention of osteoporosis, hot flashes, prevention of heart disease, insomnia, living with medical uncertainty, menstrual-type bleeding, and genitourinary symptoms.

CLINICAL QUESTION: What are the concerns of women considering hormone replacement therapy (HRT)?

BACKGROUND: Counseling women about HRT is common in family practice, but available research has focused on risks and benefits from the perspective of providers. This qualitative study explores the perspective of patients.

POPULATION STUDIED: Twenty-six women who received new HRT prescriptions at a staff-model health maintenance organization were interviewed. Originally, 176 were identified as potential participants; reasons for exclusion included premenopausal state, previous use of estrogens, and use of topical estrogens. Thirty-four women refused to participate, 20 were never contacted, and 34 consented but were never interviewed because study data were sufficient. The median age was 53, 85% were white, and median annual income was $46,000. Fifty-four percent stated they had initiated discussion of HRT with their provider, and 81% had filled the HRT prescription at the time of interview.

STUDY DESIGN AND VALIDITY: This was a qualitative study. Each woman was interviewed for 1 hour, and the audiotape was transcribed and analyzed. Three judges used a consensus process to identify the variety of domains (specific concerns) the women mentioned. Another investigator assigned patient comments to specific domains. Interviewing new subjects stopped when no new information was being added.

OUTCOMES MEASURED: This qualitative study identified womens’ concerns about HRT. By design, a qualitative study can neither quantify the prevalence or strength of particular concerns nor address clinical outcomes.

RESULTS: Influences on the decision to begin HRT were broad and included provider opinion (96% of interviewees reported), media reports (81%), experiences and opinions of friends (77%), and experiences and opinions of family (65%). Specific clinical concerns cited included risk for breast cancer, having to take medication, prevention of osteoporosis, hot flashes, prevention of heart disease, insomnia, living with medical uncertainty, menstrual-type bleeding, and genitourinary symptoms.

CLINICAL QUESTION: What are the concerns of women considering hormone replacement therapy (HRT)?

BACKGROUND: Counseling women about HRT is common in family practice, but available research has focused on risks and benefits from the perspective of providers. This qualitative study explores the perspective of patients.

POPULATION STUDIED: Twenty-six women who received new HRT prescriptions at a staff-model health maintenance organization were interviewed. Originally, 176 were identified as potential participants; reasons for exclusion included premenopausal state, previous use of estrogens, and use of topical estrogens. Thirty-four women refused to participate, 20 were never contacted, and 34 consented but were never interviewed because study data were sufficient. The median age was 53, 85% were white, and median annual income was $46,000. Fifty-four percent stated they had initiated discussion of HRT with their provider, and 81% had filled the HRT prescription at the time of interview.

STUDY DESIGN AND VALIDITY: This was a qualitative study. Each woman was interviewed for 1 hour, and the audiotape was transcribed and analyzed. Three judges used a consensus process to identify the variety of domains (specific concerns) the women mentioned. Another investigator assigned patient comments to specific domains. Interviewing new subjects stopped when no new information was being added.

OUTCOMES MEASURED: This qualitative study identified womens’ concerns about HRT. By design, a qualitative study can neither quantify the prevalence or strength of particular concerns nor address clinical outcomes.

RESULTS: Influences on the decision to begin HRT were broad and included provider opinion (96% of interviewees reported), media reports (81%), experiences and opinions of friends (77%), and experiences and opinions of family (65%). Specific clinical concerns cited included risk for breast cancer, having to take medication, prevention of osteoporosis, hot flashes, prevention of heart disease, insomnia, living with medical uncertainty, menstrual-type bleeding, and genitourinary symptoms.

CLINICAL QUESTION: Are corticosteroids effective in the treatment of croup?

BACKGROUND: Croup (laryngotracheobronchitis) is a common upper respiratory illness in children. Although it is often self-limiting, croup can cause significant morbidity and be costly to health care systems because of frequent visits to physicians or emergency departments and hospitalizations. The customary therapy of humidified air has only anecdotal benefits, and the evidence for racemic epinephrine shows only temporary relief at best.

POPULATION STUDIED: The 24 trials in this analysis included children 4 months to 12 years of age (mean ages ranged from 13 to 45 months). The authors of 14 trials studied inpatients (1375 patients), and 10 were conducted in emergency departments (846 patients).

STUDY DESIGN AND VALIDITY: This was a well-executed meta-analysis in which the authors combined the results of 24 randomized controlled trials that examined the effectiveness of corticosteroids for children with croup. Ninety-seven studies were initially identified by searching MEDLINE (1966 to 1997), EMBASE (1974 to 1997), and The Cochrane Controlled Trials Register of The Cochrane Library, as well as by corresponding with the authors of trials published in the previous 5 years. Studies were selected if 2 reviewers independently judged that these randomized controlled trials compared a corticosteroid with placebo or an active control and used a clinically important outcome. Seventy-three trials were excluded from analysis for appropriate reasons. Two independent reviewers blinded to author and study location assessed the quality of included trials using a validated 5-point scale. Systemic dexamethasone (Decadron) and nebulized budesonide (Pulmicort) were the most commonly studied drugs.

OUTCOMES MEASURED: Scores on scales measuring severity of croup symptoms, use of additional interventions, length of stay in the emergency department or hospital, and hospitalization rates were the main outcomes evaluated in this analysis. The 17-point Westley Clinical Croup Scale was the most frequently used outcome measure (13 studies). Other nonvalidated scoring systems were used in 5 studies, while 6 studies did not report a clinical croup score. Trial effect sizes were used in the pooled analyses because of inconsistency in the reporting of croup scores.

RESULTS: A significantly greater decrease in croup symptoms was noted after 6 and 12 hours of treatment with corticosteroids compared with treatment with placebo or a nonsteroid control. At 6 hours, this difference resulted in an average 15% improvement in symptom scores (95% confidence interval [CI], 2%-28%) with a number needed to treat (NNT) of 7. At 12 hours, the difference was 21% (95% CI, 9%-33%) with an NNT of 5.

CLINICAL QUESTION: Are corticosteroids effective in the treatment of croup?

BACKGROUND: Croup (laryngotracheobronchitis) is a common upper respiratory illness in children. Although it is often self-limiting, croup can cause significant morbidity and be costly to health care systems because of frequent visits to physicians or emergency departments and hospitalizations. The customary therapy of humidified air has only anecdotal benefits, and the evidence for racemic epinephrine shows only temporary relief at best.

POPULATION STUDIED: The 24 trials in this analysis included children 4 months to 12 years of age (mean ages ranged from 13 to 45 months). The authors of 14 trials studied inpatients (1375 patients), and 10 were conducted in emergency departments (846 patients).

STUDY DESIGN AND VALIDITY: This was a well-executed meta-analysis in which the authors combined the results of 24 randomized controlled trials that examined the effectiveness of corticosteroids for children with croup. Ninety-seven studies were initially identified by searching MEDLINE (1966 to 1997), EMBASE (1974 to 1997), and The Cochrane Controlled Trials Register of The Cochrane Library, as well as by corresponding with the authors of trials published in the previous 5 years. Studies were selected if 2 reviewers independently judged that these randomized controlled trials compared a corticosteroid with placebo or an active control and used a clinically important outcome. Seventy-three trials were excluded from analysis for appropriate reasons. Two independent reviewers blinded to author and study location assessed the quality of included trials using a validated 5-point scale. Systemic dexamethasone (Decadron) and nebulized budesonide (Pulmicort) were the most commonly studied drugs.

OUTCOMES MEASURED: Scores on scales measuring severity of croup symptoms, use of additional interventions, length of stay in the emergency department or hospital, and hospitalization rates were the main outcomes evaluated in this analysis. The 17-point Westley Clinical Croup Scale was the most frequently used outcome measure (13 studies). Other nonvalidated scoring systems were used in 5 studies, while 6 studies did not report a clinical croup score. Trial effect sizes were used in the pooled analyses because of inconsistency in the reporting of croup scores.

RESULTS: A significantly greater decrease in croup symptoms was noted after 6 and 12 hours of treatment with corticosteroids compared with treatment with placebo or a nonsteroid control. At 6 hours, this difference resulted in an average 15% improvement in symptom scores (95% confidence interval [CI], 2%-28%) with a number needed to treat (NNT) of 7. At 12 hours, the difference was 21% (95% CI, 9%-33%) with an NNT of 5.

CLINICAL QUESTION: Are corticosteroids effective in the treatment of croup?

BACKGROUND: Croup (laryngotracheobronchitis) is a common upper respiratory illness in children. Although it is often self-limiting, croup can cause significant morbidity and be costly to health care systems because of frequent visits to physicians or emergency departments and hospitalizations. The customary therapy of humidified air has only anecdotal benefits, and the evidence for racemic epinephrine shows only temporary relief at best.

POPULATION STUDIED: The 24 trials in this analysis included children 4 months to 12 years of age (mean ages ranged from 13 to 45 months). The authors of 14 trials studied inpatients (1375 patients), and 10 were conducted in emergency departments (846 patients).

STUDY DESIGN AND VALIDITY: This was a well-executed meta-analysis in which the authors combined the results of 24 randomized controlled trials that examined the effectiveness of corticosteroids for children with croup. Ninety-seven studies were initially identified by searching MEDLINE (1966 to 1997), EMBASE (1974 to 1997), and The Cochrane Controlled Trials Register of The Cochrane Library, as well as by corresponding with the authors of trials published in the previous 5 years. Studies were selected if 2 reviewers independently judged that these randomized controlled trials compared a corticosteroid with placebo or an active control and used a clinically important outcome. Seventy-three trials were excluded from analysis for appropriate reasons. Two independent reviewers blinded to author and study location assessed the quality of included trials using a validated 5-point scale. Systemic dexamethasone (Decadron) and nebulized budesonide (Pulmicort) were the most commonly studied drugs.

OUTCOMES MEASURED: Scores on scales measuring severity of croup symptoms, use of additional interventions, length of stay in the emergency department or hospital, and hospitalization rates were the main outcomes evaluated in this analysis. The 17-point Westley Clinical Croup Scale was the most frequently used outcome measure (13 studies). Other nonvalidated scoring systems were used in 5 studies, while 6 studies did not report a clinical croup score. Trial effect sizes were used in the pooled analyses because of inconsistency in the reporting of croup scores.

RESULTS: A significantly greater decrease in croup symptoms was noted after 6 and 12 hours of treatment with corticosteroids compared with treatment with placebo or a nonsteroid control. At 6 hours, this difference resulted in an average 15% improvement in symptom scores (95% confidence interval [CI], 2%-28%) with a number needed to treat (NNT) of 7. At 12 hours, the difference was 21% (95% CI, 9%-33%) with an NNT of 5.

CLINICAL QUESTION: Which strategy is preferable in the initial treatment of H pylori-positive patients with dyspepsia: empiric eradication therapy or endoscopy-based management?

BACKGROUND: The role of esophagogastroduodenoscopy (EGD) in the management of H pylori disease has been increasingly challenged as less invasive diagnostic tests have become available. Empiric eradication therapy is a common strategy employed in H pylori-positive dyspeptic patients. This strategy, however, has not been compared with treatment on the basis of pathologic findings using EGD in a randomized clinical trial.

POPULATION STUDIED: Study participants were all referred to specialty clinics with symptoms of ulcerlike dyspepsia. Patients aged older than 45 years, those exhibiting any alarm symptoms (eg, weight loss, history of gastrointestinal bleeding, dysphagia), and those with a recent history of treatment for H pylori were excluded. A total of 104 patients met the inclusion criteria.

STUDY DESIGN AND VALIDITY: Patients with ulcerlike dyspepsia who were positive for H pylori were randomized to either empiric eradication therapy (omeprazole, tinadazole, and clarithromycin given twice daily for 1 week) or management on the basis of findings from the EGD. Randomization was stratified to take into account risk factors including gender, tobacco use, and alcohol intake. Patients in the EGD group received eradication therapy if findings indicated peptic ulcer disease. Patients with nonulcer dyspepsia were treated symptomatically. Follow-up included a breath test for H pylori in all patients who received eradication therapy. Those patients who were still positive received second-line eradication therapy. Patients in the empiric eradication group who had no improvement in their dyspepsia scores underwent EGD even if H pylori-negative. Symptoms were reassessed at 6 weeks and 3, 6, and 12 months; H pylori status and quality of life were reassessed during the 12-month review. The study design has 3 limitations. First, since each participant did not undergo an endoscopic evaluation, it is possible that the 2 groups differed in pathology. The authors attempted to control for this by randomization and stratification of patients into risk categories. Second, the study was not blinded; participants may have been affected by the knowledge of their H pylori status or their endoscopic findings. Finally, criteria for ulcerlike dyspepsia are not clearly defined.

OUTCOMES MEASURED: The primary outcomes were a validated symptom score for dyspepsia and an overall quality-of-life score (The Medical Outcomes Study Short Form-36). Secondary outcomes included the rate of H pylori eradication and the use of symptomatic therapy at the 12-month follow-up.

RESULTS: Patients in the empiric eradication and EGD groups were similar at baseline with respect to dyspepsia scores. At 12 months, dyspepsia scores had improved in both groups, with a significantly better score in the empiric eradication group than in the EGD group (3.37 vs 5.08, P <.05). Quality-of-life scores were significantly improved in both groups compared with baseline. There was a significant difference between the 2 groups in only one parameter of quality of life; physical role functioning improved in the empiric eradication group and declined in the EGD group. Eradication rates were similar in the 2 groups (79% in the empiric eradication group vs 78% in the EGD group). More patients in the empiric eradication group were not using H2-antagonist or proton pump inhibitor therapy at 12 months (76% vs 63%). Only 27% of the patients in the empiric eradication group required EGD.

CLINICAL QUESTION: Which strategy is preferable in the initial treatment of H pylori-positive patients with dyspepsia: empiric eradication therapy or endoscopy-based management?

BACKGROUND: The role of esophagogastroduodenoscopy (EGD) in the management of H pylori disease has been increasingly challenged as less invasive diagnostic tests have become available. Empiric eradication therapy is a common strategy employed in H pylori-positive dyspeptic patients. This strategy, however, has not been compared with treatment on the basis of pathologic findings using EGD in a randomized clinical trial.

POPULATION STUDIED: Study participants were all referred to specialty clinics with symptoms of ulcerlike dyspepsia. Patients aged older than 45 years, those exhibiting any alarm symptoms (eg, weight loss, history of gastrointestinal bleeding, dysphagia), and those with a recent history of treatment for H pylori were excluded. A total of 104 patients met the inclusion criteria.

STUDY DESIGN AND VALIDITY: Patients with ulcerlike dyspepsia who were positive for H pylori were randomized to either empiric eradication therapy (omeprazole, tinadazole, and clarithromycin given twice daily for 1 week) or management on the basis of findings from the EGD. Randomization was stratified to take into account risk factors including gender, tobacco use, and alcohol intake. Patients in the EGD group received eradication therapy if findings indicated peptic ulcer disease. Patients with nonulcer dyspepsia were treated symptomatically. Follow-up included a breath test for H pylori in all patients who received eradication therapy. Those patients who were still positive received second-line eradication therapy. Patients in the empiric eradication group who had no improvement in their dyspepsia scores underwent EGD even if H pylori-negative. Symptoms were reassessed at 6 weeks and 3, 6, and 12 months; H pylori status and quality of life were reassessed during the 12-month review. The study design has 3 limitations. First, since each participant did not undergo an endoscopic evaluation, it is possible that the 2 groups differed in pathology. The authors attempted to control for this by randomization and stratification of patients into risk categories. Second, the study was not blinded; participants may have been affected by the knowledge of their H pylori status or their endoscopic findings. Finally, criteria for ulcerlike dyspepsia are not clearly defined.

OUTCOMES MEASURED: The primary outcomes were a validated symptom score for dyspepsia and an overall quality-of-life score (The Medical Outcomes Study Short Form-36). Secondary outcomes included the rate of H pylori eradication and the use of symptomatic therapy at the 12-month follow-up.

RESULTS: Patients in the empiric eradication and EGD groups were similar at baseline with respect to dyspepsia scores. At 12 months, dyspepsia scores had improved in both groups, with a significantly better score in the empiric eradication group than in the EGD group (3.37 vs 5.08, P <.05). Quality-of-life scores were significantly improved in both groups compared with baseline. There was a significant difference between the 2 groups in only one parameter of quality of life; physical role functioning improved in the empiric eradication group and declined in the EGD group. Eradication rates were similar in the 2 groups (79% in the empiric eradication group vs 78% in the EGD group). More patients in the empiric eradication group were not using H2-antagonist or proton pump inhibitor therapy at 12 months (76% vs 63%). Only 27% of the patients in the empiric eradication group required EGD.

CLINICAL QUESTION: Which strategy is preferable in the initial treatment of H pylori-positive patients with dyspepsia: empiric eradication therapy or endoscopy-based management?

BACKGROUND: The role of esophagogastroduodenoscopy (EGD) in the management of H pylori disease has been increasingly challenged as less invasive diagnostic tests have become available. Empiric eradication therapy is a common strategy employed in H pylori-positive dyspeptic patients. This strategy, however, has not been compared with treatment on the basis of pathologic findings using EGD in a randomized clinical trial.

POPULATION STUDIED: Study participants were all referred to specialty clinics with symptoms of ulcerlike dyspepsia. Patients aged older than 45 years, those exhibiting any alarm symptoms (eg, weight loss, history of gastrointestinal bleeding, dysphagia), and those with a recent history of treatment for H pylori were excluded. A total of 104 patients met the inclusion criteria.

STUDY DESIGN AND VALIDITY: Patients with ulcerlike dyspepsia who were positive for H pylori were randomized to either empiric eradication therapy (omeprazole, tinadazole, and clarithromycin given twice daily for 1 week) or management on the basis of findings from the EGD. Randomization was stratified to take into account risk factors including gender, tobacco use, and alcohol intake. Patients in the EGD group received eradication therapy if findings indicated peptic ulcer disease. Patients with nonulcer dyspepsia were treated symptomatically. Follow-up included a breath test for H pylori in all patients who received eradication therapy. Those patients who were still positive received second-line eradication therapy. Patients in the empiric eradication group who had no improvement in their dyspepsia scores underwent EGD even if H pylori-negative. Symptoms were reassessed at 6 weeks and 3, 6, and 12 months; H pylori status and quality of life were reassessed during the 12-month review. The study design has 3 limitations. First, since each participant did not undergo an endoscopic evaluation, it is possible that the 2 groups differed in pathology. The authors attempted to control for this by randomization and stratification of patients into risk categories. Second, the study was not blinded; participants may have been affected by the knowledge of their H pylori status or their endoscopic findings. Finally, criteria for ulcerlike dyspepsia are not clearly defined.

OUTCOMES MEASURED: The primary outcomes were a validated symptom score for dyspepsia and an overall quality-of-life score (The Medical Outcomes Study Short Form-36). Secondary outcomes included the rate of H pylori eradication and the use of symptomatic therapy at the 12-month follow-up.

RESULTS: Patients in the empiric eradication and EGD groups were similar at baseline with respect to dyspepsia scores. At 12 months, dyspepsia scores had improved in both groups, with a significantly better score in the empiric eradication group than in the EGD group (3.37 vs 5.08, P <.05). Quality-of-life scores were significantly improved in both groups compared with baseline. There was a significant difference between the 2 groups in only one parameter of quality of life; physical role functioning improved in the empiric eradication group and declined in the EGD group. Eradication rates were similar in the 2 groups (79% in the empiric eradication group vs 78% in the EGD group). More patients in the empiric eradication group were not using H2-antagonist or proton pump inhibitor therapy at 12 months (76% vs 63%). Only 27% of the patients in the empiric eradication group required EGD.

CLINICAL QUESTION: How effective is self-referred phone counseling for smoking cessation compared with physician advice, face-to-face nurse counseling, and phone counseling?

BACKGROUND: The Agency for Health Care Policy and Research (AHCPR) guideline on smoking cessation recommends that primary care physicians provide both brief advice and follow-up care for all smokers. Although physician advice alone achieves quit rates of less than 10%, telephone support counseling services have achieved success rates of 20% to 34%. The authors emphasize that implementing smoking cessation programs in clinical settings is difficult and not usually reimbursed. Therefore, this study addressed the question of whether smoker self-referral for phone counseling alone (community-based group) was as effective as physician advice with follow-up care consisting of both face-to-face counseling and phone counseling (practice-based group).

POPULATION STUDIED: The 319 subjects in the community-based group were recruited from mid-Michigan communities using local advertisements. The 168 subjects in the practice-based group were recruited during regular office visits to 4 university-affiliated practices in the same geographic area. All subjects were aged 18 years or older and were willing to quit smoking within 30 days. Practice-based subjects were significantly more likely to have incomes below $10,000, be covered by Medicaid, be enrolled in managed care, be less educated, and use nicotine replacement.

STUDY DESIGN AND VALIDITY: This was a nonrandomized controlled trial that compared 2 smoking cessation strategies. Nurses and phone counselors were trained in computer-assisted relapse prevention counseling and evaluated for intervention skills to improve reliability and decrease potential bias from differing counseling proficiency. Practice-based subjects received brief advice from their physicians and had 3 face-to-face sessions and 3 telephone sessions, while community-based subjects participated in 6 telephone sessions. All subjects were offered free nicotine replacement. All 6 treatment sessions were completed by 63% of practice-based subjects and 64% of community-based subjects. Confirmatory carbon monoxide monitoring was offered to every subject; however, funding limitations allowed confirmation of only 51% of self-reported smoke-free subjects. The authors attempted to control for potential self-selection bias by demonstrating that there were no significant differences between the 2 groups regarding their self-assessed confidence to quit smoking. Logistic regression analysis was also used to control for demographic differences between the 2 populations.

OUTCOMES MEASURED: The primary outcome was self-reported 7-day smoke-free status at 6 months. Secondary outcomes included: smoke-free days per participant; nurse compliance with treatment protocols; physician, nurse, and counselor satisfaction; smoker-participant satisfaction; and nicotine replacement use.

RESULTS: Using intention-to-treat analysis, 1-week abstinence rates were not statistically significant (26% and 22% for practice-based and community-based groups, respectively). One-week abstinence rates were 35% and 36%, respectively, when patients were excluded because of lack of follow-up. No significant differences were found in the logistic regression analysis. Adjusting for lack of follow-up showed no differences between the 2 groups. All participants evaluated the program positively. However, physicians and nurses felt that financial incentives and dedicated scheduling would be required to sustain a practice-based program.

CLINICAL QUESTION: How effective is self-referred phone counseling for smoking cessation compared with physician advice, face-to-face nurse counseling, and phone counseling?

BACKGROUND: The Agency for Health Care Policy and Research (AHCPR) guideline on smoking cessation recommends that primary care physicians provide both brief advice and follow-up care for all smokers. Although physician advice alone achieves quit rates of less than 10%, telephone support counseling services have achieved success rates of 20% to 34%. The authors emphasize that implementing smoking cessation programs in clinical settings is difficult and not usually reimbursed. Therefore, this study addressed the question of whether smoker self-referral for phone counseling alone (community-based group) was as effective as physician advice with follow-up care consisting of both face-to-face counseling and phone counseling (practice-based group).

POPULATION STUDIED: The 319 subjects in the community-based group were recruited from mid-Michigan communities using local advertisements. The 168 subjects in the practice-based group were recruited during regular office visits to 4 university-affiliated practices in the same geographic area. All subjects were aged 18 years or older and were willing to quit smoking within 30 days. Practice-based subjects were significantly more likely to have incomes below $10,000, be covered by Medicaid, be enrolled in managed care, be less educated, and use nicotine replacement.

STUDY DESIGN AND VALIDITY: This was a nonrandomized controlled trial that compared 2 smoking cessation strategies. Nurses and phone counselors were trained in computer-assisted relapse prevention counseling and evaluated for intervention skills to improve reliability and decrease potential bias from differing counseling proficiency. Practice-based subjects received brief advice from their physicians and had 3 face-to-face sessions and 3 telephone sessions, while community-based subjects participated in 6 telephone sessions. All subjects were offered free nicotine replacement. All 6 treatment sessions were completed by 63% of practice-based subjects and 64% of community-based subjects. Confirmatory carbon monoxide monitoring was offered to every subject; however, funding limitations allowed confirmation of only 51% of self-reported smoke-free subjects. The authors attempted to control for potential self-selection bias by demonstrating that there were no significant differences between the 2 groups regarding their self-assessed confidence to quit smoking. Logistic regression analysis was also used to control for demographic differences between the 2 populations.

OUTCOMES MEASURED: The primary outcome was self-reported 7-day smoke-free status at 6 months. Secondary outcomes included: smoke-free days per participant; nurse compliance with treatment protocols; physician, nurse, and counselor satisfaction; smoker-participant satisfaction; and nicotine replacement use.

RESULTS: Using intention-to-treat analysis, 1-week abstinence rates were not statistically significant (26% and 22% for practice-based and community-based groups, respectively). One-week abstinence rates were 35% and 36%, respectively, when patients were excluded because of lack of follow-up. No significant differences were found in the logistic regression analysis. Adjusting for lack of follow-up showed no differences between the 2 groups. All participants evaluated the program positively. However, physicians and nurses felt that financial incentives and dedicated scheduling would be required to sustain a practice-based program.

CLINICAL QUESTION: How effective is self-referred phone counseling for smoking cessation compared with physician advice, face-to-face nurse counseling, and phone counseling?

BACKGROUND: The Agency for Health Care Policy and Research (AHCPR) guideline on smoking cessation recommends that primary care physicians provide both brief advice and follow-up care for all smokers. Although physician advice alone achieves quit rates of less than 10%, telephone support counseling services have achieved success rates of 20% to 34%. The authors emphasize that implementing smoking cessation programs in clinical settings is difficult and not usually reimbursed. Therefore, this study addressed the question of whether smoker self-referral for phone counseling alone (community-based group) was as effective as physician advice with follow-up care consisting of both face-to-face counseling and phone counseling (practice-based group).

POPULATION STUDIED: The 319 subjects in the community-based group were recruited from mid-Michigan communities using local advertisements. The 168 subjects in the practice-based group were recruited during regular office visits to 4 university-affiliated practices in the same geographic area. All subjects were aged 18 years or older and were willing to quit smoking within 30 days. Practice-based subjects were significantly more likely to have incomes below $10,000, be covered by Medicaid, be enrolled in managed care, be less educated, and use nicotine replacement.

STUDY DESIGN AND VALIDITY: This was a nonrandomized controlled trial that compared 2 smoking cessation strategies. Nurses and phone counselors were trained in computer-assisted relapse prevention counseling and evaluated for intervention skills to improve reliability and decrease potential bias from differing counseling proficiency. Practice-based subjects received brief advice from their physicians and had 3 face-to-face sessions and 3 telephone sessions, while community-based subjects participated in 6 telephone sessions. All subjects were offered free nicotine replacement. All 6 treatment sessions were completed by 63% of practice-based subjects and 64% of community-based subjects. Confirmatory carbon monoxide monitoring was offered to every subject; however, funding limitations allowed confirmation of only 51% of self-reported smoke-free subjects. The authors attempted to control for potential self-selection bias by demonstrating that there were no significant differences between the 2 groups regarding their self-assessed confidence to quit smoking. Logistic regression analysis was also used to control for demographic differences between the 2 populations.

OUTCOMES MEASURED: The primary outcome was self-reported 7-day smoke-free status at 6 months. Secondary outcomes included: smoke-free days per participant; nurse compliance with treatment protocols; physician, nurse, and counselor satisfaction; smoker-participant satisfaction; and nicotine replacement use.

RESULTS: Using intention-to-treat analysis, 1-week abstinence rates were not statistically significant (26% and 22% for practice-based and community-based groups, respectively). One-week abstinence rates were 35% and 36%, respectively, when patients were excluded because of lack of follow-up. No significant differences were found in the logistic regression analysis. Adjusting for lack of follow-up showed no differences between the 2 groups. All participants evaluated the program positively. However, physicians and nurses felt that financial incentives and dedicated scheduling would be required to sustain a practice-based program.

CLINICAL QUESTION: Is estrogen replacement therapy (ERT) beneficial for women after treatment for breast cancer?

BACKGROUND: A history of breast cancer is considered a contraindication for ERT, on the assumption that this treatment promotes carcinoma of the breast and may therefore hasten recurrences and metastases. However, evidence of this is lacking. The authors of this study challenged the assumptions.

POPULATION STUDIED: All women (N = 76) with a history of breast cancer seen in an academic practice from 1978 through 1998 were included. Age ranged from 34 to 83 years with a mean of 61.8 ± 2.6 years. Information regarding ethnicity was not provided.

STUDY DESIGN AND VALIDITY: This is a retrospective case series (chart review). The methods section of the paper consists entirely of a description of the Population studied and the fact that most of the information was obtained from charts and supplemented by the hospital tumor registry. Additionally, telephone calls were made to determine the current status of those women whose records had not been updated in the past year.

OUTCOMES MEASURED: The number and percentage of each group that died during the study period and the cause of death were the main results reported. The length of time each woman was receiving ERT was a secondary measure.

RESULTS: Of the women receiving ERT, 3 (6%) died: 2 of breast cancer and 1 of myocardial infarction. This group included the one patient with advanced disease at baseline. Of the women receiving nonestrogen hormone replacement therapy, one (12.5%) died of breast cancer. Of the group receiving no hormones, 6 (33.3%) died: 5 of breast cancer and 1 of a stroke. Mean duration of ERT was 5.5 ± 2.5 years, with a range of 6 months to 32 years.

CLINICAL QUESTION: Is estrogen replacement therapy (ERT) beneficial for women after treatment for breast cancer?

BACKGROUND: A history of breast cancer is considered a contraindication for ERT, on the assumption that this treatment promotes carcinoma of the breast and may therefore hasten recurrences and metastases. However, evidence of this is lacking. The authors of this study challenged the assumptions.

POPULATION STUDIED: All women (N = 76) with a history of breast cancer seen in an academic practice from 1978 through 1998 were included. Age ranged from 34 to 83 years with a mean of 61.8 ± 2.6 years. Information regarding ethnicity was not provided.

STUDY DESIGN AND VALIDITY: This is a retrospective case series (chart review). The methods section of the paper consists entirely of a description of the Population studied and the fact that most of the information was obtained from charts and supplemented by the hospital tumor registry. Additionally, telephone calls were made to determine the current status of those women whose records had not been updated in the past year.

OUTCOMES MEASURED: The number and percentage of each group that died during the study period and the cause of death were the main results reported. The length of time each woman was receiving ERT was a secondary measure.

RESULTS: Of the women receiving ERT, 3 (6%) died: 2 of breast cancer and 1 of myocardial infarction. This group included the one patient with advanced disease at baseline. Of the women receiving nonestrogen hormone replacement therapy, one (12.5%) died of breast cancer. Of the group receiving no hormones, 6 (33.3%) died: 5 of breast cancer and 1 of a stroke. Mean duration of ERT was 5.5 ± 2.5 years, with a range of 6 months to 32 years.

CLINICAL QUESTION: Is estrogen replacement therapy (ERT) beneficial for women after treatment for breast cancer?

BACKGROUND: A history of breast cancer is considered a contraindication for ERT, on the assumption that this treatment promotes carcinoma of the breast and may therefore hasten recurrences and metastases. However, evidence of this is lacking. The authors of this study challenged the assumptions.

POPULATION STUDIED: All women (N = 76) with a history of breast cancer seen in an academic practice from 1978 through 1998 were included. Age ranged from 34 to 83 years with a mean of 61.8 ± 2.6 years. Information regarding ethnicity was not provided.

STUDY DESIGN AND VALIDITY: This is a retrospective case series (chart review). The methods section of the paper consists entirely of a description of the Population studied and the fact that most of the information was obtained from charts and supplemented by the hospital tumor registry. Additionally, telephone calls were made to determine the current status of those women whose records had not been updated in the past year.

OUTCOMES MEASURED: The number and percentage of each group that died during the study period and the cause of death were the main results reported. The length of time each woman was receiving ERT was a secondary measure.

RESULTS: Of the women receiving ERT, 3 (6%) died: 2 of breast cancer and 1 of myocardial infarction. This group included the one patient with advanced disease at baseline. Of the women receiving nonestrogen hormone replacement therapy, one (12.5%) died of breast cancer. Of the group receiving no hormones, 6 (33.3%) died: 5 of breast cancer and 1 of a stroke. Mean duration of ERT was 5.5 ± 2.5 years, with a range of 6 months to 32 years.

CLINICAL QUESTION: Is enoxaparin safe and effective for preventing venous thromboembolism (VT) and pulmonary embolism (PE) in hospitalized moderately ill nonsurgical patients?

BACKGROUND: Patients hospitalized for congestive heart failure, respiratory distress, or infection are at increased risk for venous thromboembolism (VT). Enoxaparin is effective as prophylaxis for VT in certain groups of surgical, cardiac, and stroke patients.¹ The authors of this study compared this drug in 2 different doses with placebo for the prophylaxis of VT in hospitalized medical patients.

POPULATION STUDIED: A total of 1102 patients older than 40 years from 60 centers in 9 countries were enrolled. The patients were expected to be hospitalized for at least 6 days but immobilized for not more than 3 days. Most patients were considered moderately ill and were hospitalized for congestive heart failure, respiratory failure without intubation, or infection without sepsis. Patients were excluded if they had indications for anticoagulation or if they had contraindications to anticoagulation or venography.

STUDY DESIGN AND VALIDITY: This is a randomized double-blind trial comparing placebo with 20 mg and 40 mg of enoxaparin given subcutaneously once daily. All patients were screened for deep venous thromboembolism (DVT) between days 6 and 14 with a venogram or venous ultrasound study. Patients were screened earlier if they were symptomatic for DVT or PE. Images were independently reviewed by a group of radiologists unaware of treatment assignments. Risk factors for VT or PE, the degree of illness, and the duration of hospitalization were not analyzed to determine if placebo and enoxaparin groups were statistically similar. In addition, the actual number of days of prophylaxis and the incidence of chronic anticoagulation treatment were not reported.

OUTCOMES MEASURED: The primary outcomes were the rates of DVT (symptomatic and asymptomatic) and PE (fatal and nonfatal) by day 14 as confirmed by imaging or venography. Secondary outcomes included the same results by day 110 as determined by phone follow-up (60% of patients) or visit. The rates of adverse events, such as hemorrhage and thrombocytopenia, were also reported.

RESULTS: There were 6 symptomatic DVTs, 94 asymptomatic DVTs, and 4 nonfatal PEs in the study population by day 14. The 40-mg enoxaparin group had a significantly lower overall rate of VT than the placebo group (5.5% vs 14.9%; relative risk [RR] = 0.37; 95% confidence interval [CI], 0.22-0.63; number needed to treat [NNT] = 11) as well as a lower incidence of proximal DVT (1.7% vs 4.9%, P= .04; NNT = 33). However, there was no significant decrease in the rate of symptomatic DVT (0.3%, 0.7%) or PE (0%, 1.0%). Results were similar at day 110 with a significantly lower overall rate of VT (7% vs 17.1%; RR = 0.41; 95% CI, 0.25-0.68) and a decrease in the rate of proximal DVT (2.2% vs 6.5%, P = .02). There was no significant decrease in the rates of symptomatic DVT (1.1% vs 1.5%) or PE (0% vs 1.2%). There was also no statistically significant difference between the 20-mg enoxaparin and placebo groups for any outcome at days 14 or 110 of follow-up. Of the 4 fatal PEs, 2 occurred in the 40-mg enoxaparin group 2 months after the end of treatment. By day 110, there was no significant difference between the 40-mg enoxaparin and placebo groups for the rate of death from any cause, and the rate of major hemorrhage and thrombocytopenia.

CLINICAL QUESTION: Is enoxaparin safe and effective for preventing venous thromboembolism (VT) and pulmonary embolism (PE) in hospitalized moderately ill nonsurgical patients?

BACKGROUND: Patients hospitalized for congestive heart failure, respiratory distress, or infection are at increased risk for venous thromboembolism (VT). Enoxaparin is effective as prophylaxis for VT in certain groups of surgical, cardiac, and stroke patients.¹ The authors of this study compared this drug in 2 different doses with placebo for the prophylaxis of VT in hospitalized medical patients.

POPULATION STUDIED: A total of 1102 patients older than 40 years from 60 centers in 9 countries were enrolled. The patients were expected to be hospitalized for at least 6 days but immobilized for not more than 3 days. Most patients were considered moderately ill and were hospitalized for congestive heart failure, respiratory failure without intubation, or infection without sepsis. Patients were excluded if they had indications for anticoagulation or if they had contraindications to anticoagulation or venography.

STUDY DESIGN AND VALIDITY: This is a randomized double-blind trial comparing placebo with 20 mg and 40 mg of enoxaparin given subcutaneously once daily. All patients were screened for deep venous thromboembolism (DVT) between days 6 and 14 with a venogram or venous ultrasound study. Patients were screened earlier if they were symptomatic for DVT or PE. Images were independently reviewed by a group of radiologists unaware of treatment assignments. Risk factors for VT or PE, the degree of illness, and the duration of hospitalization were not analyzed to determine if placebo and enoxaparin groups were statistically similar. In addition, the actual number of days of prophylaxis and the incidence of chronic anticoagulation treatment were not reported.

OUTCOMES MEASURED: The primary outcomes were the rates of DVT (symptomatic and asymptomatic) and PE (fatal and nonfatal) by day 14 as confirmed by imaging or venography. Secondary outcomes included the same results by day 110 as determined by phone follow-up (60% of patients) or visit. The rates of adverse events, such as hemorrhage and thrombocytopenia, were also reported.

RESULTS: There were 6 symptomatic DVTs, 94 asymptomatic DVTs, and 4 nonfatal PEs in the study population by day 14. The 40-mg enoxaparin group had a significantly lower overall rate of VT than the placebo group (5.5% vs 14.9%; relative risk [RR] = 0.37; 95% confidence interval [CI], 0.22-0.63; number needed to treat [NNT] = 11) as well as a lower incidence of proximal DVT (1.7% vs 4.9%, P= .04; NNT = 33). However, there was no significant decrease in the rate of symptomatic DVT (0.3%, 0.7%) or PE (0%, 1.0%). Results were similar at day 110 with a significantly lower overall rate of VT (7% vs 17.1%; RR = 0.41; 95% CI, 0.25-0.68) and a decrease in the rate of proximal DVT (2.2% vs 6.5%, P = .02). There was no significant decrease in the rates of symptomatic DVT (1.1% vs 1.5%) or PE (0% vs 1.2%). There was also no statistically significant difference between the 20-mg enoxaparin and placebo groups for any outcome at days 14 or 110 of follow-up. Of the 4 fatal PEs, 2 occurred in the 40-mg enoxaparin group 2 months after the end of treatment. By day 110, there was no significant difference between the 40-mg enoxaparin and placebo groups for the rate of death from any cause, and the rate of major hemorrhage and thrombocytopenia.

CLINICAL QUESTION: Is enoxaparin safe and effective for preventing venous thromboembolism (VT) and pulmonary embolism (PE) in hospitalized moderately ill nonsurgical patients?

BACKGROUND: Patients hospitalized for congestive heart failure, respiratory distress, or infection are at increased risk for venous thromboembolism (VT). Enoxaparin is effective as prophylaxis for VT in certain groups of surgical, cardiac, and stroke patients.¹ The authors of this study compared this drug in 2 different doses with placebo for the prophylaxis of VT in hospitalized medical patients.

POPULATION STUDIED: A total of 1102 patients older than 40 years from 60 centers in 9 countries were enrolled. The patients were expected to be hospitalized for at least 6 days but immobilized for not more than 3 days. Most patients were considered moderately ill and were hospitalized for congestive heart failure, respiratory failure without intubation, or infection without sepsis. Patients were excluded if they had indications for anticoagulation or if they had contraindications to anticoagulation or venography.

STUDY DESIGN AND VALIDITY: This is a randomized double-blind trial comparing placebo with 20 mg and 40 mg of enoxaparin given subcutaneously once daily. All patients were screened for deep venous thromboembolism (DVT) between days 6 and 14 with a venogram or venous ultrasound study. Patients were screened earlier if they were symptomatic for DVT or PE. Images were independently reviewed by a group of radiologists unaware of treatment assignments. Risk factors for VT or PE, the degree of illness, and the duration of hospitalization were not analyzed to determine if placebo and enoxaparin groups were statistically similar. In addition, the actual number of days of prophylaxis and the incidence of chronic anticoagulation treatment were not reported.

OUTCOMES MEASURED: The primary outcomes were the rates of DVT (symptomatic and asymptomatic) and PE (fatal and nonfatal) by day 14 as confirmed by imaging or venography. Secondary outcomes included the same results by day 110 as determined by phone follow-up (60% of patients) or visit. The rates of adverse events, such as hemorrhage and thrombocytopenia, were also reported.

RESULTS: There were 6 symptomatic DVTs, 94 asymptomatic DVTs, and 4 nonfatal PEs in the study population by day 14. The 40-mg enoxaparin group had a significantly lower overall rate of VT than the placebo group (5.5% vs 14.9%; relative risk [RR] = 0.37; 95% confidence interval [CI], 0.22-0.63; number needed to treat [NNT] = 11) as well as a lower incidence of proximal DVT (1.7% vs 4.9%, P= .04; NNT = 33). However, there was no significant decrease in the rate of symptomatic DVT (0.3%, 0.7%) or PE (0%, 1.0%). Results were similar at day 110 with a significantly lower overall rate of VT (7% vs 17.1%; RR = 0.41; 95% CI, 0.25-0.68) and a decrease in the rate of proximal DVT (2.2% vs 6.5%, P = .02). There was no significant decrease in the rates of symptomatic DVT (1.1% vs 1.5%) or PE (0% vs 1.2%). There was also no statistically significant difference between the 20-mg enoxaparin and placebo groups for any outcome at days 14 or 110 of follow-up. Of the 4 fatal PEs, 2 occurred in the 40-mg enoxaparin group 2 months after the end of treatment. By day 110, there was no significant difference between the 40-mg enoxaparin and placebo groups for the rate of death from any cause, and the rate of major hemorrhage and thrombocytopenia.

CLINICAL QUESTION: Is Echinacea beneficial for the prevention or treatment of acute upper respiratory infections (URIs)?

BACKGROUND: Echinacea, taken for the prevention and treatment of URIs, is one of the most commonly used herbs. The goal of the authors of this study was to systematically review the randomized trials of Echinacea for the prevention or treatment of acute URIs.

POPULATION STUDIED: The authors reviewed the evidence from blinded placebo-controlled randomized trials of any formulation of Echinacea used for prevention or treatment of URIs. The study populations varied, and most were in Europe.

STUDY DESIGN AND VALIDITY: The authors identified studies by searching MEDLINE and other bibliographic reference services using variants on the term “Echinacea.” They also reviewed articles, books, and book chapters for references. They questioned herbal medicine experts in the United States and Germany about published and unpublished trials. Study quality was evaluated using the following criteria: randomization, blinding, power, validity, clinical relevance of outcome measurements, inclusion and exclusion criteria, indistinguishability between treatment and placebo, and appropriateness of conclusions. The authors went to great lengths to identify relevant studies, and it seems unlikely that any pertinent studies would have been missed. The authors do not describe how the quality considerations were scored or if they were reproducible. They state that meta-analysis (using statistical methods to combine data from different studies into a single summary measure of effect) was not an appropriate option because of the variations in preparations, methods, and outcomes measured. Even without meta-analysis, this is a good systematic review of the evidence for Echinacea for the prevention and treatment of URIs.

OUTCOMES MEASURED: The outcomes used in the treatment studies varied considerably. They included URI symptoms rated on a scale of 0 to 3, progression to a “real cold,” and the presence of influenzalike symptoms. Outcomes for the prevention trials included the time until the first URI and the severity of its symptoms.

RESULTS: The authors identified 13 blinded randomized trials. Nine of these studies were treatment trials, and 4 were studies of prevention. Eight of the 9 treatment trials reported a benefit. Six showed a statistically significant benefit, while 2 reported only a trend; the one study showing no benefit is unpublished and used insufficient doses of Echinacea. Data from one of the larger more methodologically sound studies¹ suggest that the number needed to treat to prevent a real cold among patients with the first sign of a cold was 5. Two of the prevention trials reported a marginal benefit: One initially reported a benefit but later reported no benefit, and one found no benefit. The major methodologic weaknesses identified included: lack of objective validated measures, no clear evidence that the treatment was indistinguishable from the placebo, and insufficient sample size.

CLINICAL QUESTION: Is Echinacea beneficial for the prevention or treatment of acute upper respiratory infections (URIs)?

BACKGROUND: Echinacea, taken for the prevention and treatment of URIs, is one of the most commonly used herbs. The goal of the authors of this study was to systematically review the randomized trials of Echinacea for the prevention or treatment of acute URIs.

POPULATION STUDIED: The authors reviewed the evidence from blinded placebo-controlled randomized trials of any formulation of Echinacea used for prevention or treatment of URIs. The study populations varied, and most were in Europe.

STUDY DESIGN AND VALIDITY: The authors identified studies by searching MEDLINE and other bibliographic reference services using variants on the term “Echinacea.” They also reviewed articles, books, and book chapters for references. They questioned herbal medicine experts in the United States and Germany about published and unpublished trials. Study quality was evaluated using the following criteria: randomization, blinding, power, validity, clinical relevance of outcome measurements, inclusion and exclusion criteria, indistinguishability between treatment and placebo, and appropriateness of conclusions. The authors went to great lengths to identify relevant studies, and it seems unlikely that any pertinent studies would have been missed. The authors do not describe how the quality considerations were scored or if they were reproducible. They state that meta-analysis (using statistical methods to combine data from different studies into a single summary measure of effect) was not an appropriate option because of the variations in preparations, methods, and outcomes measured. Even without meta-analysis, this is a good systematic review of the evidence for Echinacea for the prevention and treatment of URIs.

OUTCOMES MEASURED: The outcomes used in the treatment studies varied considerably. They included URI symptoms rated on a scale of 0 to 3, progression to a “real cold,” and the presence of influenzalike symptoms. Outcomes for the prevention trials included the time until the first URI and the severity of its symptoms.

RESULTS: The authors identified 13 blinded randomized trials. Nine of these studies were treatment trials, and 4 were studies of prevention. Eight of the 9 treatment trials reported a benefit. Six showed a statistically significant benefit, while 2 reported only a trend; the one study showing no benefit is unpublished and used insufficient doses of Echinacea. Data from one of the larger more methodologically sound studies¹ suggest that the number needed to treat to prevent a real cold among patients with the first sign of a cold was 5. Two of the prevention trials reported a marginal benefit: One initially reported a benefit but later reported no benefit, and one found no benefit. The major methodologic weaknesses identified included: lack of objective validated measures, no clear evidence that the treatment was indistinguishable from the placebo, and insufficient sample size.

CLINICAL QUESTION: Is Echinacea beneficial for the prevention or treatment of acute upper respiratory infections (URIs)?

BACKGROUND: Echinacea, taken for the prevention and treatment of URIs, is one of the most commonly used herbs. The goal of the authors of this study was to systematically review the randomized trials of Echinacea for the prevention or treatment of acute URIs.

POPULATION STUDIED: The authors reviewed the evidence from blinded placebo-controlled randomized trials of any formulation of Echinacea used for prevention or treatment of URIs. The study populations varied, and most were in Europe.

STUDY DESIGN AND VALIDITY: The authors identified studies by searching MEDLINE and other bibliographic reference services using variants on the term “Echinacea.” They also reviewed articles, books, and book chapters for references. They questioned herbal medicine experts in the United States and Germany about published and unpublished trials. Study quality was evaluated using the following criteria: randomization, blinding, power, validity, clinical relevance of outcome measurements, inclusion and exclusion criteria, indistinguishability between treatment and placebo, and appropriateness of conclusions. The authors went to great lengths to identify relevant studies, and it seems unlikely that any pertinent studies would have been missed. The authors do not describe how the quality considerations were scored or if they were reproducible. They state that meta-analysis (using statistical methods to combine data from different studies into a single summary measure of effect) was not an appropriate option because of the variations in preparations, methods, and outcomes measured. Even without meta-analysis, this is a good systematic review of the evidence for Echinacea for the prevention and treatment of URIs.

OUTCOMES MEASURED: The outcomes used in the treatment studies varied considerably. They included URI symptoms rated on a scale of 0 to 3, progression to a “real cold,” and the presence of influenzalike symptoms. Outcomes for the prevention trials included the time until the first URI and the severity of its symptoms.

RESULTS: The authors identified 13 blinded randomized trials. Nine of these studies were treatment trials, and 4 were studies of prevention. Eight of the 9 treatment trials reported a benefit. Six showed a statistically significant benefit, while 2 reported only a trend; the one study showing no benefit is unpublished and used insufficient doses of Echinacea. Data from one of the larger more methodologically sound studies¹ suggest that the number needed to treat to prevent a real cold among patients with the first sign of a cold was 5. Two of the prevention trials reported a marginal benefit: One initially reported a benefit but later reported no benefit, and one found no benefit. The major methodologic weaknesses identified included: lack of objective validated measures, no clear evidence that the treatment was indistinguishable from the placebo, and insufficient sample size.