User login
Which venous leg ulcers will heal using limb compression bandages?
BACKGROUND: Limb compression bandages, such as Unna boots, successfully treat 30% to 60% of venous leg ulcers. The authors of this study attempted to create a simple model to predict which patients’ ulcers will heal using this therapy.
POPULATION STUDIED: All patients were drawn from a dermatology skin ulcer clinic. A training group of 260 consecutive patients with chronic venous leg ulcers was used to develop predictive models. Venous ulcers were defined as wounds in the “gaiter area” of the limb from the mid-calf to 1 inch below the malleolus. All patients had lower leg edema and cutaneous evidence of venous disease. second group of 219 patients enrolled in the control arm of 2 randomized controlled trials testing skin allograft versus compression was used to validate the rule. Ulcers in this study were at least 4 weeks old and between 0.5 and 200 cm2 in area.
STUDY DESIGN AND VALIDITY: This was a cohort study for developing and validating a clinical decision rule. Prognostic models were created with 10 risk factors and statistically analyzed for their ability to correctly predict wound healing. The models were simplified by dropping prognostic factors one by one from the model, as long as the ability to predict healing was not affected. All of the models were evaluated by measuring the area under the receiver-operating characteristic (ROC) curve. By definition, an area closer to 1 indicates a better ability to distinguish which ulcers will heal. An area of 0.5 indicates a worthless test or rule. The most accurate model from the training group was evaluated using the validation group.
OUTCOMES MEASURED: The primary outcome measure was the ability of a model to predict wound healing within 24 weeks as measured by the area under the ROC curve.
RESULTS: Overall, 65% of ulcers from the training group and 56% of ulcers from the validation group healed by 24 weeks. All of the candidate models had areas under the ROC curve greater than 0.80, meaning that the model discriminated between ulcers that did and did not heal more than 80% of the time. The simplest model, which gave a score of 1 point for a wound area greater than 5 cm2 (measured by a template) and 1 point for wound age greater than 6 months, correctly predicted healing 87% of the time. Ulcers with a score of 0 healed in 95% of the validation group, while ulcers with a score of 1 healed in 73% of the validation group.
This simple rule accurately predicts which venous stasis ulcers will heal in 6 months using compression bandages. Score 1 point for a wound older than 6 months and 1 point if the area of the wound is greater than 5 cm2 as measured by a template. Ulcers with a score of 0 are 95% likely to heal, while those with a score of 2 are only 37% likely to heal. Patients whose ulcers rate a score of 0 can be treated with compression bandages and are unlikely to need referral. Whether newer therapies will add benefit for patients with a score of 1 or 2 is unknown. Early referral of patients with ulcers with a score of 2 will likely depend on local practice.
BACKGROUND: Limb compression bandages, such as Unna boots, successfully treat 30% to 60% of venous leg ulcers. The authors of this study attempted to create a simple model to predict which patients’ ulcers will heal using this therapy.
POPULATION STUDIED: All patients were drawn from a dermatology skin ulcer clinic. A training group of 260 consecutive patients with chronic venous leg ulcers was used to develop predictive models. Venous ulcers were defined as wounds in the “gaiter area” of the limb from the mid-calf to 1 inch below the malleolus. All patients had lower leg edema and cutaneous evidence of venous disease. second group of 219 patients enrolled in the control arm of 2 randomized controlled trials testing skin allograft versus compression was used to validate the rule. Ulcers in this study were at least 4 weeks old and between 0.5 and 200 cm2 in area.
STUDY DESIGN AND VALIDITY: This was a cohort study for developing and validating a clinical decision rule. Prognostic models were created with 10 risk factors and statistically analyzed for their ability to correctly predict wound healing. The models were simplified by dropping prognostic factors one by one from the model, as long as the ability to predict healing was not affected. All of the models were evaluated by measuring the area under the receiver-operating characteristic (ROC) curve. By definition, an area closer to 1 indicates a better ability to distinguish which ulcers will heal. An area of 0.5 indicates a worthless test or rule. The most accurate model from the training group was evaluated using the validation group.
OUTCOMES MEASURED: The primary outcome measure was the ability of a model to predict wound healing within 24 weeks as measured by the area under the ROC curve.
RESULTS: Overall, 65% of ulcers from the training group and 56% of ulcers from the validation group healed by 24 weeks. All of the candidate models had areas under the ROC curve greater than 0.80, meaning that the model discriminated between ulcers that did and did not heal more than 80% of the time. The simplest model, which gave a score of 1 point for a wound area greater than 5 cm2 (measured by a template) and 1 point for wound age greater than 6 months, correctly predicted healing 87% of the time. Ulcers with a score of 0 healed in 95% of the validation group, while ulcers with a score of 1 healed in 73% of the validation group.
This simple rule accurately predicts which venous stasis ulcers will heal in 6 months using compression bandages. Score 1 point for a wound older than 6 months and 1 point if the area of the wound is greater than 5 cm2 as measured by a template. Ulcers with a score of 0 are 95% likely to heal, while those with a score of 2 are only 37% likely to heal. Patients whose ulcers rate a score of 0 can be treated with compression bandages and are unlikely to need referral. Whether newer therapies will add benefit for patients with a score of 1 or 2 is unknown. Early referral of patients with ulcers with a score of 2 will likely depend on local practice.
BACKGROUND: Limb compression bandages, such as Unna boots, successfully treat 30% to 60% of venous leg ulcers. The authors of this study attempted to create a simple model to predict which patients’ ulcers will heal using this therapy.
POPULATION STUDIED: All patients were drawn from a dermatology skin ulcer clinic. A training group of 260 consecutive patients with chronic venous leg ulcers was used to develop predictive models. Venous ulcers were defined as wounds in the “gaiter area” of the limb from the mid-calf to 1 inch below the malleolus. All patients had lower leg edema and cutaneous evidence of venous disease. second group of 219 patients enrolled in the control arm of 2 randomized controlled trials testing skin allograft versus compression was used to validate the rule. Ulcers in this study were at least 4 weeks old and between 0.5 and 200 cm2 in area.
STUDY DESIGN AND VALIDITY: This was a cohort study for developing and validating a clinical decision rule. Prognostic models were created with 10 risk factors and statistically analyzed for their ability to correctly predict wound healing. The models were simplified by dropping prognostic factors one by one from the model, as long as the ability to predict healing was not affected. All of the models were evaluated by measuring the area under the receiver-operating characteristic (ROC) curve. By definition, an area closer to 1 indicates a better ability to distinguish which ulcers will heal. An area of 0.5 indicates a worthless test or rule. The most accurate model from the training group was evaluated using the validation group.
OUTCOMES MEASURED: The primary outcome measure was the ability of a model to predict wound healing within 24 weeks as measured by the area under the ROC curve.
RESULTS: Overall, 65% of ulcers from the training group and 56% of ulcers from the validation group healed by 24 weeks. All of the candidate models had areas under the ROC curve greater than 0.80, meaning that the model discriminated between ulcers that did and did not heal more than 80% of the time. The simplest model, which gave a score of 1 point for a wound area greater than 5 cm2 (measured by a template) and 1 point for wound age greater than 6 months, correctly predicted healing 87% of the time. Ulcers with a score of 0 healed in 95% of the validation group, while ulcers with a score of 1 healed in 73% of the validation group.
This simple rule accurately predicts which venous stasis ulcers will heal in 6 months using compression bandages. Score 1 point for a wound older than 6 months and 1 point if the area of the wound is greater than 5 cm2 as measured by a template. Ulcers with a score of 0 are 95% likely to heal, while those with a score of 2 are only 37% likely to heal. Patients whose ulcers rate a score of 0 can be treated with compression bandages and are unlikely to need referral. Whether newer therapies will add benefit for patients with a score of 1 or 2 is unknown. Early referral of patients with ulcers with a score of 2 will likely depend on local practice.
Which patients with an ischemic stroke are likely to be helped or harmed by early aspirin treatment?
BACKGROUND: Long-term aspirin use is of demonstrated benefit in the prevention of ischemic stroke and death in high-risk patients. The Chinese Acute Stroke Trial (CAST) and the International Stroke Trial (IST) showed that beginning aspirin promptly after an ischemic stroke reduces the immediate risk of recurrent stroke and death. The authors of this paper pooled the results of these 2 trials to evaluate outcomes in different subgroups.
POPULATION STUDIED: The IST included 19,435 patients from 36 countries across Europe, Asia, North America, and South America, while the CAST reported on 20,655 patients from China. All of the patients presented within 48 hours of symptom onset (or sleep onset if the stroke occurred while asleep). The patients were included in both studies if their physician determined that there were no contraindications to aspirin treatment but he or she was uncertain whether aspirin treatment was indicated. Patients with a clear reason to receive aspirin were excluded, such as patients with possible cardiac ischemia, as were those with a contraindication to aspirin treatment. Contraindications were determined by the responsible physician rather than a protocol and were generally based on the likelihood of an increased risk of adverse effects (eg, aspirin sensitivity, already orally anticoagulated, or recent serious gastric bleeding) or the likelihood of little benefit (eg, severe preexisting disability or symptoms likely to quickly resolve). The mean age of patients was 67 years; 12% had atrial fibrillation, and 15% had a systolic blood pressure of at least 190 mm Hg
STUDY DESIGN AND VALIDITY: The CAST used concealed randomization, blinding, and placebo control. Although most patients had computed tomography (CT) scanning before randomization, only comatose patients were required to have CT scans to exclude intracranial hemorrhage. Aspirin (160 mg film-coated) or placebo was given for 4 weeks or until earlier hospital discharge or death. In the CAST trial, 219 patients in the aspirin group and 232 in the placebo group lacked follow-up data and were excluded from the analysis. The IST used concealed randomization but lacked placebo control. Aspirin in a 300-mg dose was given for 2 weeks or until earlier hospital discharge or death. Half of the patients in the IST also received subcutaneous heparin. Only 2 patients were lost to follow-up in the IST. Heterogeneity between studies was appropriately evaluated and found to be nonsignificant. The Mantel-Haenszel chi-square method was used to combine data from the studies, which is appropriate when the studies are homogenous. No racial subgroups were analyzed, and it would be helpful to confirm that this benefit holds true across these subgroups.
OUTCOMES MEASURED: The primary outcomes were the rates of recurrent ischemic stroke, hemorraghic stroke, any stroke, death without further stroke, or the combined outcome (any of these). These outcomes were presented for different patient subgroups based on the hours since stroke onset, age, sex, level of consciousness, presence of atrial fibrillation, whether CT was performed, CT findings, initial systolic blood pressure, stroke syndrome (lacunar vs nonlacunar), score on a prognostic index, and whether the patient received heparin.
RESULTS: The rates of recurrent ischemic stroke (1.6% vs 2.3%; P <.001; number needed to treat [NNT]=143) and death without further stroke (5.0% vs 5.4%; P=.05; NNT=250) were both significantly lower in the aspirin treatment group. The risk of hemorrhagic stroke or hemorrhagic transformation of the original stroke was not significantly increased by use of aspirin (1.0% vs 0.8%; P=.07%). There was a significant decrease in the combined outcome of further stroke or death (8.2% vs 9.1%; P=.001; NNT=111). The subgroup analysis did not reveal any characteristic that defined a group that varied significantly in the degree of benefit or harm from the entire treated group.
These large studies demonstrate that aspirin (160 to 300 mg per day for 2 to 4 weeks) is effective in the treatment of acute ischemic stroke. The NNT of 111 to prevent 1 stroke or death is large, but aspirin is inexpensive and easy to administer. The study did not reveal any subgroup that did better or worse than the overall group. Therefore, if there is no contraindication to aspirin use in the judgment of the treating physician, it can be started immediately in the patient with acute ischemic stroke. One area of uncertainty not addressed by this study is the issue of concomitant use of aspirin and thrombolytics, which was shown to increase the risk of death over the use of thrombolytics alone in another study.1
BACKGROUND: Long-term aspirin use is of demonstrated benefit in the prevention of ischemic stroke and death in high-risk patients. The Chinese Acute Stroke Trial (CAST) and the International Stroke Trial (IST) showed that beginning aspirin promptly after an ischemic stroke reduces the immediate risk of recurrent stroke and death. The authors of this paper pooled the results of these 2 trials to evaluate outcomes in different subgroups.
POPULATION STUDIED: The IST included 19,435 patients from 36 countries across Europe, Asia, North America, and South America, while the CAST reported on 20,655 patients from China. All of the patients presented within 48 hours of symptom onset (or sleep onset if the stroke occurred while asleep). The patients were included in both studies if their physician determined that there were no contraindications to aspirin treatment but he or she was uncertain whether aspirin treatment was indicated. Patients with a clear reason to receive aspirin were excluded, such as patients with possible cardiac ischemia, as were those with a contraindication to aspirin treatment. Contraindications were determined by the responsible physician rather than a protocol and were generally based on the likelihood of an increased risk of adverse effects (eg, aspirin sensitivity, already orally anticoagulated, or recent serious gastric bleeding) or the likelihood of little benefit (eg, severe preexisting disability or symptoms likely to quickly resolve). The mean age of patients was 67 years; 12% had atrial fibrillation, and 15% had a systolic blood pressure of at least 190 mm Hg
STUDY DESIGN AND VALIDITY: The CAST used concealed randomization, blinding, and placebo control. Although most patients had computed tomography (CT) scanning before randomization, only comatose patients were required to have CT scans to exclude intracranial hemorrhage. Aspirin (160 mg film-coated) or placebo was given for 4 weeks or until earlier hospital discharge or death. In the CAST trial, 219 patients in the aspirin group and 232 in the placebo group lacked follow-up data and were excluded from the analysis. The IST used concealed randomization but lacked placebo control. Aspirin in a 300-mg dose was given for 2 weeks or until earlier hospital discharge or death. Half of the patients in the IST also received subcutaneous heparin. Only 2 patients were lost to follow-up in the IST. Heterogeneity between studies was appropriately evaluated and found to be nonsignificant. The Mantel-Haenszel chi-square method was used to combine data from the studies, which is appropriate when the studies are homogenous. No racial subgroups were analyzed, and it would be helpful to confirm that this benefit holds true across these subgroups.
OUTCOMES MEASURED: The primary outcomes were the rates of recurrent ischemic stroke, hemorraghic stroke, any stroke, death without further stroke, or the combined outcome (any of these). These outcomes were presented for different patient subgroups based on the hours since stroke onset, age, sex, level of consciousness, presence of atrial fibrillation, whether CT was performed, CT findings, initial systolic blood pressure, stroke syndrome (lacunar vs nonlacunar), score on a prognostic index, and whether the patient received heparin.
RESULTS: The rates of recurrent ischemic stroke (1.6% vs 2.3%; P <.001; number needed to treat [NNT]=143) and death without further stroke (5.0% vs 5.4%; P=.05; NNT=250) were both significantly lower in the aspirin treatment group. The risk of hemorrhagic stroke or hemorrhagic transformation of the original stroke was not significantly increased by use of aspirin (1.0% vs 0.8%; P=.07%). There was a significant decrease in the combined outcome of further stroke or death (8.2% vs 9.1%; P=.001; NNT=111). The subgroup analysis did not reveal any characteristic that defined a group that varied significantly in the degree of benefit or harm from the entire treated group.
These large studies demonstrate that aspirin (160 to 300 mg per day for 2 to 4 weeks) is effective in the treatment of acute ischemic stroke. The NNT of 111 to prevent 1 stroke or death is large, but aspirin is inexpensive and easy to administer. The study did not reveal any subgroup that did better or worse than the overall group. Therefore, if there is no contraindication to aspirin use in the judgment of the treating physician, it can be started immediately in the patient with acute ischemic stroke. One area of uncertainty not addressed by this study is the issue of concomitant use of aspirin and thrombolytics, which was shown to increase the risk of death over the use of thrombolytics alone in another study.1
BACKGROUND: Long-term aspirin use is of demonstrated benefit in the prevention of ischemic stroke and death in high-risk patients. The Chinese Acute Stroke Trial (CAST) and the International Stroke Trial (IST) showed that beginning aspirin promptly after an ischemic stroke reduces the immediate risk of recurrent stroke and death. The authors of this paper pooled the results of these 2 trials to evaluate outcomes in different subgroups.
POPULATION STUDIED: The IST included 19,435 patients from 36 countries across Europe, Asia, North America, and South America, while the CAST reported on 20,655 patients from China. All of the patients presented within 48 hours of symptom onset (or sleep onset if the stroke occurred while asleep). The patients were included in both studies if their physician determined that there were no contraindications to aspirin treatment but he or she was uncertain whether aspirin treatment was indicated. Patients with a clear reason to receive aspirin were excluded, such as patients with possible cardiac ischemia, as were those with a contraindication to aspirin treatment. Contraindications were determined by the responsible physician rather than a protocol and were generally based on the likelihood of an increased risk of adverse effects (eg, aspirin sensitivity, already orally anticoagulated, or recent serious gastric bleeding) or the likelihood of little benefit (eg, severe preexisting disability or symptoms likely to quickly resolve). The mean age of patients was 67 years; 12% had atrial fibrillation, and 15% had a systolic blood pressure of at least 190 mm Hg
STUDY DESIGN AND VALIDITY: The CAST used concealed randomization, blinding, and placebo control. Although most patients had computed tomography (CT) scanning before randomization, only comatose patients were required to have CT scans to exclude intracranial hemorrhage. Aspirin (160 mg film-coated) or placebo was given for 4 weeks or until earlier hospital discharge or death. In the CAST trial, 219 patients in the aspirin group and 232 in the placebo group lacked follow-up data and were excluded from the analysis. The IST used concealed randomization but lacked placebo control. Aspirin in a 300-mg dose was given for 2 weeks or until earlier hospital discharge or death. Half of the patients in the IST also received subcutaneous heparin. Only 2 patients were lost to follow-up in the IST. Heterogeneity between studies was appropriately evaluated and found to be nonsignificant. The Mantel-Haenszel chi-square method was used to combine data from the studies, which is appropriate when the studies are homogenous. No racial subgroups were analyzed, and it would be helpful to confirm that this benefit holds true across these subgroups.
OUTCOMES MEASURED: The primary outcomes were the rates of recurrent ischemic stroke, hemorraghic stroke, any stroke, death without further stroke, or the combined outcome (any of these). These outcomes were presented for different patient subgroups based on the hours since stroke onset, age, sex, level of consciousness, presence of atrial fibrillation, whether CT was performed, CT findings, initial systolic blood pressure, stroke syndrome (lacunar vs nonlacunar), score on a prognostic index, and whether the patient received heparin.
RESULTS: The rates of recurrent ischemic stroke (1.6% vs 2.3%; P <.001; number needed to treat [NNT]=143) and death without further stroke (5.0% vs 5.4%; P=.05; NNT=250) were both significantly lower in the aspirin treatment group. The risk of hemorrhagic stroke or hemorrhagic transformation of the original stroke was not significantly increased by use of aspirin (1.0% vs 0.8%; P=.07%). There was a significant decrease in the combined outcome of further stroke or death (8.2% vs 9.1%; P=.001; NNT=111). The subgroup analysis did not reveal any characteristic that defined a group that varied significantly in the degree of benefit or harm from the entire treated group.
These large studies demonstrate that aspirin (160 to 300 mg per day for 2 to 4 weeks) is effective in the treatment of acute ischemic stroke. The NNT of 111 to prevent 1 stroke or death is large, but aspirin is inexpensive and easy to administer. The study did not reveal any subgroup that did better or worse than the overall group. Therefore, if there is no contraindication to aspirin use in the judgment of the treating physician, it can be started immediately in the patient with acute ischemic stroke. One area of uncertainty not addressed by this study is the issue of concomitant use of aspirin and thrombolytics, which was shown to increase the risk of death over the use of thrombolytics alone in another study.1
What is the risk of advanced proximal colon neoplasms in patients with and without distal colon disease?
BACKGROUND: Flexible sigmoidoscopy is the standard screening test for colon cancer in asymptomatic individuals older than 50 years who do not have risk factors. It is assumed that a normal examination of the distal colon is associated with a reduced risk of proximal colon neoplasia. The authors of this study attempted to quantify the relative risk of advanced proximal colon neoplasia in patients with and without distal polyps.
POPULATION STUDIED: Employees of a pharmaceutical company and their family members who were 50 years or older were given the opportunity for free colonoscopic screening as a health benefit. The average age of the 1994 participants was 59.8 years, and 58.9% were men. Patients with a previous personal history of colorectal polyps, colorectal cancer, or inflammatory bowel disease and those with symptoms such as rectal bleeding, change in bowel habits, or recent abdominal pain were excluded.
STUDY DESIGN AND VALIDITY: This is a cross-sectional analysis of patients who received colonoscopy between 1995 and 1998. Patients were given information regarding the purpose of the study and were asked to call if interested in participating. A telephone interview was conducted to determine eligibility. Colonoscopy was performed by 36 gastroenterologists, and 97% of examinations reached the cecum. The size and location of polyps identified on colonoscopy were recorded before removal. Pathologists from the same center categorized and characterized the specimens. A limitation of the study is the absence of information about family history, particularly since patients with a positive family history may have been more likely to volunteer for participation.
OUTCOMES MEASURED: The main outcome of this study was the relative risk (RR) of advanced colonic neoplasia proximal to the splenic flexure in patients with and without distal colonic lesions. An advanced distal or proximal lesion was defined as a polyp or polypoid lesion with villous features, high-grade dysplasia, or cancer.
RESULTS: Of the 1994 patients studied, 61 (3.1%) had advanced distal neoplasia and 50 (2.5%) had advanced proximal lesions. This included 5 distal cancers and 7 proximal cancers. In patients with no distal lesions (78% of the population), 1.5% had an advanced proximal lesion. This was the reference group for the groups with pathological findings in the distal colon. The prevalence of advanced proximal lesions was 4.0% (RR=2.6; 95% confidence interval [CI], 1.1-5.9) in the group with distal hyperplastic polyps (10.1% of the population), 7.1% (RR=4.0; 95% CI, 1.9-8.3) in the group with distal tubular adenoma (8.4% of the population), and 11.5% (RR=6.7; 95% CI, 3.2-16.6) in the group with advanced distal neoplasms (3.1% of the population). Although the RR of advanced proximal neoplasia increased with the severity of distal lesions as expected, 46% of all advanced proximal neoplasms occurred in subjects with no distal lesions.
The risk of advanced proximal neoplasia is related to the severity of findings in the distal colon. Another study in the same issue1 also showed that a significant number of advanced proximal colon lesions go unrecognized with a standard flexible sigmoidoscopy investigation, even when combined with fecal occult blood testing. In fact, approximately half of the patients with advanced proximal colon neoplasia had no evidence of distal lesions. Although this study adds to a body of evidence that suggests that colonoscopy may have a role in screening for colon cancer, guidelines are needed that take into consideration not only these data but also cost, availability, and the impact of different screening modalities on mortality.
BACKGROUND: Flexible sigmoidoscopy is the standard screening test for colon cancer in asymptomatic individuals older than 50 years who do not have risk factors. It is assumed that a normal examination of the distal colon is associated with a reduced risk of proximal colon neoplasia. The authors of this study attempted to quantify the relative risk of advanced proximal colon neoplasia in patients with and without distal polyps.
POPULATION STUDIED: Employees of a pharmaceutical company and their family members who were 50 years or older were given the opportunity for free colonoscopic screening as a health benefit. The average age of the 1994 participants was 59.8 years, and 58.9% were men. Patients with a previous personal history of colorectal polyps, colorectal cancer, or inflammatory bowel disease and those with symptoms such as rectal bleeding, change in bowel habits, or recent abdominal pain were excluded.
STUDY DESIGN AND VALIDITY: This is a cross-sectional analysis of patients who received colonoscopy between 1995 and 1998. Patients were given information regarding the purpose of the study and were asked to call if interested in participating. A telephone interview was conducted to determine eligibility. Colonoscopy was performed by 36 gastroenterologists, and 97% of examinations reached the cecum. The size and location of polyps identified on colonoscopy were recorded before removal. Pathologists from the same center categorized and characterized the specimens. A limitation of the study is the absence of information about family history, particularly since patients with a positive family history may have been more likely to volunteer for participation.
OUTCOMES MEASURED: The main outcome of this study was the relative risk (RR) of advanced colonic neoplasia proximal to the splenic flexure in patients with and without distal colonic lesions. An advanced distal or proximal lesion was defined as a polyp or polypoid lesion with villous features, high-grade dysplasia, or cancer.
RESULTS: Of the 1994 patients studied, 61 (3.1%) had advanced distal neoplasia and 50 (2.5%) had advanced proximal lesions. This included 5 distal cancers and 7 proximal cancers. In patients with no distal lesions (78% of the population), 1.5% had an advanced proximal lesion. This was the reference group for the groups with pathological findings in the distal colon. The prevalence of advanced proximal lesions was 4.0% (RR=2.6; 95% confidence interval [CI], 1.1-5.9) in the group with distal hyperplastic polyps (10.1% of the population), 7.1% (RR=4.0; 95% CI, 1.9-8.3) in the group with distal tubular adenoma (8.4% of the population), and 11.5% (RR=6.7; 95% CI, 3.2-16.6) in the group with advanced distal neoplasms (3.1% of the population). Although the RR of advanced proximal neoplasia increased with the severity of distal lesions as expected, 46% of all advanced proximal neoplasms occurred in subjects with no distal lesions.
The risk of advanced proximal neoplasia is related to the severity of findings in the distal colon. Another study in the same issue1 also showed that a significant number of advanced proximal colon lesions go unrecognized with a standard flexible sigmoidoscopy investigation, even when combined with fecal occult blood testing. In fact, approximately half of the patients with advanced proximal colon neoplasia had no evidence of distal lesions. Although this study adds to a body of evidence that suggests that colonoscopy may have a role in screening for colon cancer, guidelines are needed that take into consideration not only these data but also cost, availability, and the impact of different screening modalities on mortality.
BACKGROUND: Flexible sigmoidoscopy is the standard screening test for colon cancer in asymptomatic individuals older than 50 years who do not have risk factors. It is assumed that a normal examination of the distal colon is associated with a reduced risk of proximal colon neoplasia. The authors of this study attempted to quantify the relative risk of advanced proximal colon neoplasia in patients with and without distal polyps.
POPULATION STUDIED: Employees of a pharmaceutical company and their family members who were 50 years or older were given the opportunity for free colonoscopic screening as a health benefit. The average age of the 1994 participants was 59.8 years, and 58.9% were men. Patients with a previous personal history of colorectal polyps, colorectal cancer, or inflammatory bowel disease and those with symptoms such as rectal bleeding, change in bowel habits, or recent abdominal pain were excluded.
STUDY DESIGN AND VALIDITY: This is a cross-sectional analysis of patients who received colonoscopy between 1995 and 1998. Patients were given information regarding the purpose of the study and were asked to call if interested in participating. A telephone interview was conducted to determine eligibility. Colonoscopy was performed by 36 gastroenterologists, and 97% of examinations reached the cecum. The size and location of polyps identified on colonoscopy were recorded before removal. Pathologists from the same center categorized and characterized the specimens. A limitation of the study is the absence of information about family history, particularly since patients with a positive family history may have been more likely to volunteer for participation.
OUTCOMES MEASURED: The main outcome of this study was the relative risk (RR) of advanced colonic neoplasia proximal to the splenic flexure in patients with and without distal colonic lesions. An advanced distal or proximal lesion was defined as a polyp or polypoid lesion with villous features, high-grade dysplasia, or cancer.
RESULTS: Of the 1994 patients studied, 61 (3.1%) had advanced distal neoplasia and 50 (2.5%) had advanced proximal lesions. This included 5 distal cancers and 7 proximal cancers. In patients with no distal lesions (78% of the population), 1.5% had an advanced proximal lesion. This was the reference group for the groups with pathological findings in the distal colon. The prevalence of advanced proximal lesions was 4.0% (RR=2.6; 95% confidence interval [CI], 1.1-5.9) in the group with distal hyperplastic polyps (10.1% of the population), 7.1% (RR=4.0; 95% CI, 1.9-8.3) in the group with distal tubular adenoma (8.4% of the population), and 11.5% (RR=6.7; 95% CI, 3.2-16.6) in the group with advanced distal neoplasms (3.1% of the population). Although the RR of advanced proximal neoplasia increased with the severity of distal lesions as expected, 46% of all advanced proximal neoplasms occurred in subjects with no distal lesions.
The risk of advanced proximal neoplasia is related to the severity of findings in the distal colon. Another study in the same issue1 also showed that a significant number of advanced proximal colon lesions go unrecognized with a standard flexible sigmoidoscopy investigation, even when combined with fecal occult blood testing. In fact, approximately half of the patients with advanced proximal colon neoplasia had no evidence of distal lesions. Although this study adds to a body of evidence that suggests that colonoscopy may have a role in screening for colon cancer, guidelines are needed that take into consideration not only these data but also cost, availability, and the impact of different screening modalities on mortality.
How accurate are rapid polymerase chain reaction tests in detecting group B streptococcus colonization in pregnant women?
BACKGROUND: Current recommendations for reducing the neonatal morbidity and mortality caused by group B streptococcus (GBS) infections support either screening pregnant women at 35 to 37 weeks’ gestation and treating those with positive cultures or treating GBS during labor on the basis of risk factors such as fever or prolonged rupture of membranes. Both of these strategies result in overtreatment (GBS-negative mothers receiving antibiotics) and undertreatment (GBS-positive mothers not receiving antibiotics). A rapid test performed and available at the time of labor may be a useful guide to appropriately treating at-risk mothers, if such a test is sufficiently accurate.
POPULATION STUDIED: A total of 112 pregnant women hospitalized for delivery in Quebec City, Canada, were enrolled. Of these, 57 had intact membranes, and 55 had membranes already ruptured. The authors did not comment on how the women were recruited or enrolled in the study. The prevalence of GBS colonization in this study population was 29.5%, which is similar to the prevalence reported in other community-based studies. The results are therefore likely to be representative of those patients seen by family physicians practicing obstetrics.
STUDY DESIGN AND VALIDITY: Investigators obtained rectal, vaginal, and rectovaginal swabs from the women at the time of admission and again after the rupture of membranes (for those whose membranes were intact on admission). Three tests were performed on each swab at the same laboratory: a standard GBS culture, a conventional polymerase chain reaction (PCR) assay, and a new rapid PCR assay. It was not stated whether those performing the PCR assays were blinded to the culture results.
OUTCOMES MEASURED: PCR assay accuracy was reported in terms of sensitivity, specificity, and predictive values, using the GBS culture as the gold standard.
RESULTS: Using the combined vaginal and anal swab before the rupture of membranes (N=57), each PCR test had a sensitivity and specificity of 100%. For samples obtained after the membranes had ruptured, the sensitivity dropped to only 93.8% for each test (only one false-negative result), while the specificity remained 100%. This corresponds to a positive predictive value of 100% and a negative predictive value of 98.8%. The positive likelihood ratio calculated from this data is 194; the negative likelihood ratio is 0.03. Results were available for the new PCR assay in 45 minutes, for the conventional PCR assay in 100 minutes, and for the GBS culture in 36 hours.
This study shows that PCR assays are highly sensitive and specific in detecting GBS carrier status in pregnant women at the time of delivery and are more rapid than the standard GBS culture. Assuming this test is available in a clinical setting and can be reliably performed, the PCR assay serves as an alternative approach to the management of GBS in pregnancy, especially in hospitals where the screen-and-treat approach is preferred and GBS status is unknown at the time of labor. A rapid test will not suffice in cases of precipitous delivery, which occurred in 15% of the study institution’s deliveries. Implementation studies in other settings are needed to determine if PCR testing is more cost-effective and whether it should replace a clinician’s current approach.
BACKGROUND: Current recommendations for reducing the neonatal morbidity and mortality caused by group B streptococcus (GBS) infections support either screening pregnant women at 35 to 37 weeks’ gestation and treating those with positive cultures or treating GBS during labor on the basis of risk factors such as fever or prolonged rupture of membranes. Both of these strategies result in overtreatment (GBS-negative mothers receiving antibiotics) and undertreatment (GBS-positive mothers not receiving antibiotics). A rapid test performed and available at the time of labor may be a useful guide to appropriately treating at-risk mothers, if such a test is sufficiently accurate.
POPULATION STUDIED: A total of 112 pregnant women hospitalized for delivery in Quebec City, Canada, were enrolled. Of these, 57 had intact membranes, and 55 had membranes already ruptured. The authors did not comment on how the women were recruited or enrolled in the study. The prevalence of GBS colonization in this study population was 29.5%, which is similar to the prevalence reported in other community-based studies. The results are therefore likely to be representative of those patients seen by family physicians practicing obstetrics.
STUDY DESIGN AND VALIDITY: Investigators obtained rectal, vaginal, and rectovaginal swabs from the women at the time of admission and again after the rupture of membranes (for those whose membranes were intact on admission). Three tests were performed on each swab at the same laboratory: a standard GBS culture, a conventional polymerase chain reaction (PCR) assay, and a new rapid PCR assay. It was not stated whether those performing the PCR assays were blinded to the culture results.
OUTCOMES MEASURED: PCR assay accuracy was reported in terms of sensitivity, specificity, and predictive values, using the GBS culture as the gold standard.
RESULTS: Using the combined vaginal and anal swab before the rupture of membranes (N=57), each PCR test had a sensitivity and specificity of 100%. For samples obtained after the membranes had ruptured, the sensitivity dropped to only 93.8% for each test (only one false-negative result), while the specificity remained 100%. This corresponds to a positive predictive value of 100% and a negative predictive value of 98.8%. The positive likelihood ratio calculated from this data is 194; the negative likelihood ratio is 0.03. Results were available for the new PCR assay in 45 minutes, for the conventional PCR assay in 100 minutes, and for the GBS culture in 36 hours.
This study shows that PCR assays are highly sensitive and specific in detecting GBS carrier status in pregnant women at the time of delivery and are more rapid than the standard GBS culture. Assuming this test is available in a clinical setting and can be reliably performed, the PCR assay serves as an alternative approach to the management of GBS in pregnancy, especially in hospitals where the screen-and-treat approach is preferred and GBS status is unknown at the time of labor. A rapid test will not suffice in cases of precipitous delivery, which occurred in 15% of the study institution’s deliveries. Implementation studies in other settings are needed to determine if PCR testing is more cost-effective and whether it should replace a clinician’s current approach.
BACKGROUND: Current recommendations for reducing the neonatal morbidity and mortality caused by group B streptococcus (GBS) infections support either screening pregnant women at 35 to 37 weeks’ gestation and treating those with positive cultures or treating GBS during labor on the basis of risk factors such as fever or prolonged rupture of membranes. Both of these strategies result in overtreatment (GBS-negative mothers receiving antibiotics) and undertreatment (GBS-positive mothers not receiving antibiotics). A rapid test performed and available at the time of labor may be a useful guide to appropriately treating at-risk mothers, if such a test is sufficiently accurate.
POPULATION STUDIED: A total of 112 pregnant women hospitalized for delivery in Quebec City, Canada, were enrolled. Of these, 57 had intact membranes, and 55 had membranes already ruptured. The authors did not comment on how the women were recruited or enrolled in the study. The prevalence of GBS colonization in this study population was 29.5%, which is similar to the prevalence reported in other community-based studies. The results are therefore likely to be representative of those patients seen by family physicians practicing obstetrics.
STUDY DESIGN AND VALIDITY: Investigators obtained rectal, vaginal, and rectovaginal swabs from the women at the time of admission and again after the rupture of membranes (for those whose membranes were intact on admission). Three tests were performed on each swab at the same laboratory: a standard GBS culture, a conventional polymerase chain reaction (PCR) assay, and a new rapid PCR assay. It was not stated whether those performing the PCR assays were blinded to the culture results.
OUTCOMES MEASURED: PCR assay accuracy was reported in terms of sensitivity, specificity, and predictive values, using the GBS culture as the gold standard.
RESULTS: Using the combined vaginal and anal swab before the rupture of membranes (N=57), each PCR test had a sensitivity and specificity of 100%. For samples obtained after the membranes had ruptured, the sensitivity dropped to only 93.8% for each test (only one false-negative result), while the specificity remained 100%. This corresponds to a positive predictive value of 100% and a negative predictive value of 98.8%. The positive likelihood ratio calculated from this data is 194; the negative likelihood ratio is 0.03. Results were available for the new PCR assay in 45 minutes, for the conventional PCR assay in 100 minutes, and for the GBS culture in 36 hours.
This study shows that PCR assays are highly sensitive and specific in detecting GBS carrier status in pregnant women at the time of delivery and are more rapid than the standard GBS culture. Assuming this test is available in a clinical setting and can be reliably performed, the PCR assay serves as an alternative approach to the management of GBS in pregnancy, especially in hospitals where the screen-and-treat approach is preferred and GBS status is unknown at the time of labor. A rapid test will not suffice in cases of precipitous delivery, which occurred in 15% of the study institution’s deliveries. Implementation studies in other settings are needed to determine if PCR testing is more cost-effective and whether it should replace a clinician’s current approach.
The Journal of Family Practice ©2000 Dowden Health Media
Which class of antidepressants is most effective and best tolerated by patients with major depression?
BACKGROUND: In 1993 the Agency for Health Care Policy and Research published guidelines for the treatment of major depression that focused on the efficacy of older antidepressants, particularly second- and third-generation tricyclics and monoamine oxidase inhibitors. Since that time several newer antidepressants have been developed. This report provides further information by including data on newer agents marketed for the treatment of depression.
POPULATION STUDIED: This meta-analysis evaluated more than 300 randomized controlled trials of newer pharmacotherapies for depression, including the selective serotonin reuptake inhibitors (SSRIs) venlafaxine, mirtazapine, nefazodone, and bupropion. Approximately 160 trials were based in outpatient practices. Trials were included if they lasted at least 6 weeks and compared a newer antidepressant with either an older or newer antidepressant or placebo. The most severely depressed patients were often excluded from the studies
STUDY DESIGN AND VALIDITY: Trials for inclusion in the analysis were identified from the Cochrane registry, reference lists from trial articles and meta-analyses, and experts. The Cochrane registry includes trials identified from multiple sources such as MEDLINE, EMBASE, and PsycLIT. A separate search identified more than 200 trials that assessed adverse drug effects. Two trained reviewers independently examined data regarding participant and diagnostic descriptors, intervention characteristics, study designs, and clinical outcomes. Disagreements were resolved by consensus. Publication bias was addressed by using funnel plots and did not suggest missing studies comparing newer antidepressants with older antidepressants. However, an overestimation of treatment effect was suggested for newer antidepressants compared with placebo. Strengths of this meta-analysis include the thorough search and the quality of the review process. A weakness of this report lies not with the meta-analysis itself but in the inconsistency and infrequency of the reporting of the data in the trials. In addition, it is limited by the lack of comparative studies of several of the newer antidepressants with other newer or older antidepressants.
OUTCOME MEASURED: Symptomatic response rate, total discontinuation rates, and discontinuation rates due to adverse events were the primary outcomes. Response rates defined as a 50% or greater improvement in symptoms were assessed by a depression symptoms rating scale or a clinical global assessment rating scale and were calculated using a modified intention-to-treat approach. This approach calculates the response rate as the number of patients who continue treatment and improve divided by the total number of patients randomized. Because some patients who left treatment may have responded, this approach provides a conservative estimate of treatment effect.
RESULTS: Response rates were significantly higher for newer pharmacotherapies than for placebo (50% vs 32%, number needed to treat=6). The new agents were as effective as the first- and second-generation tricyclic antidepressants. No difference was found among the different classes of newer antidepressants, though comparative data were few and use short-term (6-week) outcomes. Fluoxetine and sertraline were effective for treating adults with dysthymia, but information was not available on other newer antidepressants. Hypericum (St John’s wort) was more effective than placebo and similar to low-dose tricyclic antidepressants for short-term treatment of mild to moderate depression. With regard to side effects, diarrhea, nausea, insomnia, and headache were associated with SSRIs, while tricyclics caused the well-known effects of dry mouth, constipation, dizziness, blurred vision, and tremors. Data were insufficient to determine the effects on sexual dysfunction and the common adverse effects of other newer therapies. Overall discontinuation rates due to adverse effects were similar for newer and older antidepressants (4% vs 5%).
The newer antidepressants are as effective as older antidepressants for the treatment of major depression and dysthymia in adults. However, data are insufficient to evaluate the effectiveness of newer therapies in other depressive disorders and populations, such as children and adolescents or postpartum women. Adverse effect profiles differ among the classes of antidepressants, but discontinuation rates are similar. These results are the same as those reported by the clinical guideline on the treatment of acute major depression of the American College of Physicians.1 When choosing treatment, clinicians should consider comorbid disease states and medications, dosing schedule, and cost, and discuss with patients the different side effect profiles of the various antidepressants to determine their preference.
BACKGROUND: In 1993 the Agency for Health Care Policy and Research published guidelines for the treatment of major depression that focused on the efficacy of older antidepressants, particularly second- and third-generation tricyclics and monoamine oxidase inhibitors. Since that time several newer antidepressants have been developed. This report provides further information by including data on newer agents marketed for the treatment of depression.
POPULATION STUDIED: This meta-analysis evaluated more than 300 randomized controlled trials of newer pharmacotherapies for depression, including the selective serotonin reuptake inhibitors (SSRIs) venlafaxine, mirtazapine, nefazodone, and bupropion. Approximately 160 trials were based in outpatient practices. Trials were included if they lasted at least 6 weeks and compared a newer antidepressant with either an older or newer antidepressant or placebo. The most severely depressed patients were often excluded from the studies
STUDY DESIGN AND VALIDITY: Trials for inclusion in the analysis were identified from the Cochrane registry, reference lists from trial articles and meta-analyses, and experts. The Cochrane registry includes trials identified from multiple sources such as MEDLINE, EMBASE, and PsycLIT. A separate search identified more than 200 trials that assessed adverse drug effects. Two trained reviewers independently examined data regarding participant and diagnostic descriptors, intervention characteristics, study designs, and clinical outcomes. Disagreements were resolved by consensus. Publication bias was addressed by using funnel plots and did not suggest missing studies comparing newer antidepressants with older antidepressants. However, an overestimation of treatment effect was suggested for newer antidepressants compared with placebo. Strengths of this meta-analysis include the thorough search and the quality of the review process. A weakness of this report lies not with the meta-analysis itself but in the inconsistency and infrequency of the reporting of the data in the trials. In addition, it is limited by the lack of comparative studies of several of the newer antidepressants with other newer or older antidepressants.
OUTCOME MEASURED: Symptomatic response rate, total discontinuation rates, and discontinuation rates due to adverse events were the primary outcomes. Response rates defined as a 50% or greater improvement in symptoms were assessed by a depression symptoms rating scale or a clinical global assessment rating scale and were calculated using a modified intention-to-treat approach. This approach calculates the response rate as the number of patients who continue treatment and improve divided by the total number of patients randomized. Because some patients who left treatment may have responded, this approach provides a conservative estimate of treatment effect.
RESULTS: Response rates were significantly higher for newer pharmacotherapies than for placebo (50% vs 32%, number needed to treat=6). The new agents were as effective as the first- and second-generation tricyclic antidepressants. No difference was found among the different classes of newer antidepressants, though comparative data were few and use short-term (6-week) outcomes. Fluoxetine and sertraline were effective for treating adults with dysthymia, but information was not available on other newer antidepressants. Hypericum (St John’s wort) was more effective than placebo and similar to low-dose tricyclic antidepressants for short-term treatment of mild to moderate depression. With regard to side effects, diarrhea, nausea, insomnia, and headache were associated with SSRIs, while tricyclics caused the well-known effects of dry mouth, constipation, dizziness, blurred vision, and tremors. Data were insufficient to determine the effects on sexual dysfunction and the common adverse effects of other newer therapies. Overall discontinuation rates due to adverse effects were similar for newer and older antidepressants (4% vs 5%).
The newer antidepressants are as effective as older antidepressants for the treatment of major depression and dysthymia in adults. However, data are insufficient to evaluate the effectiveness of newer therapies in other depressive disorders and populations, such as children and adolescents or postpartum women. Adverse effect profiles differ among the classes of antidepressants, but discontinuation rates are similar. These results are the same as those reported by the clinical guideline on the treatment of acute major depression of the American College of Physicians.1 When choosing treatment, clinicians should consider comorbid disease states and medications, dosing schedule, and cost, and discuss with patients the different side effect profiles of the various antidepressants to determine their preference.
BACKGROUND: In 1993 the Agency for Health Care Policy and Research published guidelines for the treatment of major depression that focused on the efficacy of older antidepressants, particularly second- and third-generation tricyclics and monoamine oxidase inhibitors. Since that time several newer antidepressants have been developed. This report provides further information by including data on newer agents marketed for the treatment of depression.
POPULATION STUDIED: This meta-analysis evaluated more than 300 randomized controlled trials of newer pharmacotherapies for depression, including the selective serotonin reuptake inhibitors (SSRIs) venlafaxine, mirtazapine, nefazodone, and bupropion. Approximately 160 trials were based in outpatient practices. Trials were included if they lasted at least 6 weeks and compared a newer antidepressant with either an older or newer antidepressant or placebo. The most severely depressed patients were often excluded from the studies
STUDY DESIGN AND VALIDITY: Trials for inclusion in the analysis were identified from the Cochrane registry, reference lists from trial articles and meta-analyses, and experts. The Cochrane registry includes trials identified from multiple sources such as MEDLINE, EMBASE, and PsycLIT. A separate search identified more than 200 trials that assessed adverse drug effects. Two trained reviewers independently examined data regarding participant and diagnostic descriptors, intervention characteristics, study designs, and clinical outcomes. Disagreements were resolved by consensus. Publication bias was addressed by using funnel plots and did not suggest missing studies comparing newer antidepressants with older antidepressants. However, an overestimation of treatment effect was suggested for newer antidepressants compared with placebo. Strengths of this meta-analysis include the thorough search and the quality of the review process. A weakness of this report lies not with the meta-analysis itself but in the inconsistency and infrequency of the reporting of the data in the trials. In addition, it is limited by the lack of comparative studies of several of the newer antidepressants with other newer or older antidepressants.
OUTCOME MEASURED: Symptomatic response rate, total discontinuation rates, and discontinuation rates due to adverse events were the primary outcomes. Response rates defined as a 50% or greater improvement in symptoms were assessed by a depression symptoms rating scale or a clinical global assessment rating scale and were calculated using a modified intention-to-treat approach. This approach calculates the response rate as the number of patients who continue treatment and improve divided by the total number of patients randomized. Because some patients who left treatment may have responded, this approach provides a conservative estimate of treatment effect.
RESULTS: Response rates were significantly higher for newer pharmacotherapies than for placebo (50% vs 32%, number needed to treat=6). The new agents were as effective as the first- and second-generation tricyclic antidepressants. No difference was found among the different classes of newer antidepressants, though comparative data were few and use short-term (6-week) outcomes. Fluoxetine and sertraline were effective for treating adults with dysthymia, but information was not available on other newer antidepressants. Hypericum (St John’s wort) was more effective than placebo and similar to low-dose tricyclic antidepressants for short-term treatment of mild to moderate depression. With regard to side effects, diarrhea, nausea, insomnia, and headache were associated with SSRIs, while tricyclics caused the well-known effects of dry mouth, constipation, dizziness, blurred vision, and tremors. Data were insufficient to determine the effects on sexual dysfunction and the common adverse effects of other newer therapies. Overall discontinuation rates due to adverse effects were similar for newer and older antidepressants (4% vs 5%).
The newer antidepressants are as effective as older antidepressants for the treatment of major depression and dysthymia in adults. However, data are insufficient to evaluate the effectiveness of newer therapies in other depressive disorders and populations, such as children and adolescents or postpartum women. Adverse effect profiles differ among the classes of antidepressants, but discontinuation rates are similar. These results are the same as those reported by the clinical guideline on the treatment of acute major depression of the American College of Physicians.1 When choosing treatment, clinicians should consider comorbid disease states and medications, dosing schedule, and cost, and discuss with patients the different side effect profiles of the various antidepressants to determine their preference.
Which patients with minor head injury do not need computed tomography?
BACKGROUND: Every year more than 2 million adults present to emergency departments in the United States with acute head trauma. Only 6% to 9% of those with apparently minor injury have an intracranial lesion by computed tomography (CT), and less than 1% will require neurosurgical intervention.1 This study develops and validates a clinical prediction guide for identifying patients with minor head injury who do not require a head CT.
POPULATION STUDIED: The study took place at a large inner city level 1 trauma center. All patients were aged at least 3 years and presented within 24 hours of a minor head injury. Minor head injury was defined as any loss of consciousness (witnessed, reported, or unknown) with a normal neurological examination and a Glasgow Coma Scale score of 15 with or without isolated deficits in short-term memory. Patients were excluded if they refused CT, had concurrent injuries precluding CT, or reported no loss of consciousness. The mean age was 36 years, and 65% were men.
STUDY DESIGN AND VALIDITY: This is a prospective validation of a clinical prediction guide. In phase 1 demographic data, symptoms, and physical findings were recorded for 520 consecutive patients with minor head injury. All of the study patients subsequently underwent CT of the head. Clinical criteria were correlated with CT findings, and a set of 7 findings was found that identified all patients with a positive CT. In phase 2, this set of 7 findings was prospectively validated in a group of 909 consecutive patients with minor head injury. The patients were analyzed in 2 groups: those who had at least one of the 7 findings and those who had none. Emergency medicine residents performed the clinical evaluations under faculty supervision. A second physician repeated the evaluation for 50 patients with 92% agreement (k=0.78). Staff neuroradiologists interpreted CT results, although it is a limitation of the study that the radiologists were apparently not blinded to the clinical condition of patients. An independent staff radiologist who was unaware of the original interpretations reviewed 50 randomly selected CT scans with 98% agreement (k=0.94). The rate of positive CT in each phase (6.9% and 6.3%, respectively) was consistent with previous studies.
OUTCOMES MEASURED: The primary outcome was the negative predictive value of the clinical prediction guide for the presence of significant head injury.
RESULTS: The following 7 findings identified all patients with a positive CT: (1) posttraumatic head pain, (2) posttraumatic emesis, (3) older than 60 years, (4) drug or alcohol intoxication, (5) deficits in short-term memory, (6) physical evidence of trauma above the clavicles, and (7) seizure. In phase 2 of the study the absence of all 7 of these findings had a negative predictive value for a positive CT of 100% (95% confidence interval, 99%-100%). Application of this clinical prediction guide to the 909 patients in phase 2 would have reduced the use of head CT by 22% without missing any abnormalities. This study did not have the power to evaluate the association seen in previous studies between coagulopathy and a positive head CT.
The authors of this study present compelling evidence that head CT is not needed for minor head injury in patients with none of the 7 findings listed. One unnecessary CT would be prevented for every 5 patients evaluated using this clinical decision rule. Based on previous studies, physicians should also obtain a CT on any patient with a coagulopathy. Given the potential for harm if the rule is inaccurate, it should be validated in another study before widespread use.
BACKGROUND: Every year more than 2 million adults present to emergency departments in the United States with acute head trauma. Only 6% to 9% of those with apparently minor injury have an intracranial lesion by computed tomography (CT), and less than 1% will require neurosurgical intervention.1 This study develops and validates a clinical prediction guide for identifying patients with minor head injury who do not require a head CT.
POPULATION STUDIED: The study took place at a large inner city level 1 trauma center. All patients were aged at least 3 years and presented within 24 hours of a minor head injury. Minor head injury was defined as any loss of consciousness (witnessed, reported, or unknown) with a normal neurological examination and a Glasgow Coma Scale score of 15 with or without isolated deficits in short-term memory. Patients were excluded if they refused CT, had concurrent injuries precluding CT, or reported no loss of consciousness. The mean age was 36 years, and 65% were men.
STUDY DESIGN AND VALIDITY: This is a prospective validation of a clinical prediction guide. In phase 1 demographic data, symptoms, and physical findings were recorded for 520 consecutive patients with minor head injury. All of the study patients subsequently underwent CT of the head. Clinical criteria were correlated with CT findings, and a set of 7 findings was found that identified all patients with a positive CT. In phase 2, this set of 7 findings was prospectively validated in a group of 909 consecutive patients with minor head injury. The patients were analyzed in 2 groups: those who had at least one of the 7 findings and those who had none. Emergency medicine residents performed the clinical evaluations under faculty supervision. A second physician repeated the evaluation for 50 patients with 92% agreement (k=0.78). Staff neuroradiologists interpreted CT results, although it is a limitation of the study that the radiologists were apparently not blinded to the clinical condition of patients. An independent staff radiologist who was unaware of the original interpretations reviewed 50 randomly selected CT scans with 98% agreement (k=0.94). The rate of positive CT in each phase (6.9% and 6.3%, respectively) was consistent with previous studies.
OUTCOMES MEASURED: The primary outcome was the negative predictive value of the clinical prediction guide for the presence of significant head injury.
RESULTS: The following 7 findings identified all patients with a positive CT: (1) posttraumatic head pain, (2) posttraumatic emesis, (3) older than 60 years, (4) drug or alcohol intoxication, (5) deficits in short-term memory, (6) physical evidence of trauma above the clavicles, and (7) seizure. In phase 2 of the study the absence of all 7 of these findings had a negative predictive value for a positive CT of 100% (95% confidence interval, 99%-100%). Application of this clinical prediction guide to the 909 patients in phase 2 would have reduced the use of head CT by 22% without missing any abnormalities. This study did not have the power to evaluate the association seen in previous studies between coagulopathy and a positive head CT.
The authors of this study present compelling evidence that head CT is not needed for minor head injury in patients with none of the 7 findings listed. One unnecessary CT would be prevented for every 5 patients evaluated using this clinical decision rule. Based on previous studies, physicians should also obtain a CT on any patient with a coagulopathy. Given the potential for harm if the rule is inaccurate, it should be validated in another study before widespread use.
BACKGROUND: Every year more than 2 million adults present to emergency departments in the United States with acute head trauma. Only 6% to 9% of those with apparently minor injury have an intracranial lesion by computed tomography (CT), and less than 1% will require neurosurgical intervention.1 This study develops and validates a clinical prediction guide for identifying patients with minor head injury who do not require a head CT.
POPULATION STUDIED: The study took place at a large inner city level 1 trauma center. All patients were aged at least 3 years and presented within 24 hours of a minor head injury. Minor head injury was defined as any loss of consciousness (witnessed, reported, or unknown) with a normal neurological examination and a Glasgow Coma Scale score of 15 with or without isolated deficits in short-term memory. Patients were excluded if they refused CT, had concurrent injuries precluding CT, or reported no loss of consciousness. The mean age was 36 years, and 65% were men.
STUDY DESIGN AND VALIDITY: This is a prospective validation of a clinical prediction guide. In phase 1 demographic data, symptoms, and physical findings were recorded for 520 consecutive patients with minor head injury. All of the study patients subsequently underwent CT of the head. Clinical criteria were correlated with CT findings, and a set of 7 findings was found that identified all patients with a positive CT. In phase 2, this set of 7 findings was prospectively validated in a group of 909 consecutive patients with minor head injury. The patients were analyzed in 2 groups: those who had at least one of the 7 findings and those who had none. Emergency medicine residents performed the clinical evaluations under faculty supervision. A second physician repeated the evaluation for 50 patients with 92% agreement (k=0.78). Staff neuroradiologists interpreted CT results, although it is a limitation of the study that the radiologists were apparently not blinded to the clinical condition of patients. An independent staff radiologist who was unaware of the original interpretations reviewed 50 randomly selected CT scans with 98% agreement (k=0.94). The rate of positive CT in each phase (6.9% and 6.3%, respectively) was consistent with previous studies.
OUTCOMES MEASURED: The primary outcome was the negative predictive value of the clinical prediction guide for the presence of significant head injury.
RESULTS: The following 7 findings identified all patients with a positive CT: (1) posttraumatic head pain, (2) posttraumatic emesis, (3) older than 60 years, (4) drug or alcohol intoxication, (5) deficits in short-term memory, (6) physical evidence of trauma above the clavicles, and (7) seizure. In phase 2 of the study the absence of all 7 of these findings had a negative predictive value for a positive CT of 100% (95% confidence interval, 99%-100%). Application of this clinical prediction guide to the 909 patients in phase 2 would have reduced the use of head CT by 22% without missing any abnormalities. This study did not have the power to evaluate the association seen in previous studies between coagulopathy and a positive head CT.
The authors of this study present compelling evidence that head CT is not needed for minor head injury in patients with none of the 7 findings listed. One unnecessary CT would be prevented for every 5 patients evaluated using this clinical decision rule. Based on previous studies, physicians should also obtain a CT on any patient with a coagulopathy. Given the potential for harm if the rule is inaccurate, it should be validated in another study before widespread use.
What is the optimal treatment for lateral ankle ligament ruptures?
BACKGROUND: The nearly 23,000 daily ankle injuries in the United States account for approximately 25% of all musculoskeletal injuries. The anterior talofibular ligament (a lateral supporting ligament) is most often involved. Current management options include casting, surgery, functional treatment, or minimal-to-no treatment. Despite recently published reviews suggesting that functional treatment and early mobilization provide the best results, other studies recommend operative treatment. This meta-analysis was designed to systematically determine the best treatment approach.
POPULATION STUDIED: More than 3300 patients with acute lateral ankle ligament ruptures enrolled in 27 randomized controlled trials were evaluated. Fifteen of the 42 trials initially identified were excluded because of poor follow-up (<40% dropout), republication of the same trial, or lack of relevant outcome variables. Diagnosis based on imaging by an arthrogram or stress radiograph was required. Study duration ranged from 6 to 46 months. Articles were included regardless of published language. Patients were largely recruited from the general population, so the results likely apply to those managed by family physicians.
STUDY DESIGN AND VALIDITY: The authors of this meta-analysis reviewed randomized controlled trials found by searching the MEDLINE, Cochrane, and EMBASE databases from 1966 to 1998. Three independent investigators blinded to author and institution reviewed each trial and assessed its quality. Quality was determined by randomization that included concealed allocation assignment, adequate follow-up, and blinded outcome assessment. Treatments included operative (using a similar surgical protocol), casting for 6 weeks, and different types of functional intervention (bracing, wrapping, orthoses such as air casts, special shoes, and cast immobilization for 3 weeks). Three methods employing minimal treatment (less than 3 weeks of socks, braces, wraps, or casts) were also included because no trials of nonintervention were identified. Data were analyzed for homogeneity by calculating the Q statistic.
OUTCOMES MEASURED: The authors evaluated 2 long-term outcomes (residual pain and giving way) and one short-term outcome (time lost from work).
RESULTS: The natural history of untreated lateral ankle ligament ruptures is unknown. Treatments of short duration or inadequate ankle joint support are associated with more frequent residual symptoms (pain and giving way). Operative treatment followed by functional treatment resulted in less giving way than functional treatment alone (relative risk [RR]=0.23; 95% confidence interval [CI], 0.17-0.31). Functional treatment alone led to less giving way than 6-week cast treatment (RR=0.69; 95% CI, 0.50-0.94). Functional treatment was also associated with less pain than 6-week cast treatment (RR=0.67; 95% CI, 0.50-0.90). No other differences between interventions were found with regard to pain. Time lost from work varied greatly. Approximately 10% of patients undergoing operative treatment developed surgical complications. Results of a subgroup analysis of the high-quality studies only were not different from the results of the corresponding analysis of all studies.
To improve long-term ankle stability, patients with lateral ligament ruptures should be managed aggressively by either functional treatment or surgery followed by functional treatment. Because surgery is costly and associated with possible complications, a trial of functional treatment should be attempted first. If an initial nonsurgical approach fails, subsequent operative reconstruction is warranted.
BACKGROUND: The nearly 23,000 daily ankle injuries in the United States account for approximately 25% of all musculoskeletal injuries. The anterior talofibular ligament (a lateral supporting ligament) is most often involved. Current management options include casting, surgery, functional treatment, or minimal-to-no treatment. Despite recently published reviews suggesting that functional treatment and early mobilization provide the best results, other studies recommend operative treatment. This meta-analysis was designed to systematically determine the best treatment approach.
POPULATION STUDIED: More than 3300 patients with acute lateral ankle ligament ruptures enrolled in 27 randomized controlled trials were evaluated. Fifteen of the 42 trials initially identified were excluded because of poor follow-up (<40% dropout), republication of the same trial, or lack of relevant outcome variables. Diagnosis based on imaging by an arthrogram or stress radiograph was required. Study duration ranged from 6 to 46 months. Articles were included regardless of published language. Patients were largely recruited from the general population, so the results likely apply to those managed by family physicians.
STUDY DESIGN AND VALIDITY: The authors of this meta-analysis reviewed randomized controlled trials found by searching the MEDLINE, Cochrane, and EMBASE databases from 1966 to 1998. Three independent investigators blinded to author and institution reviewed each trial and assessed its quality. Quality was determined by randomization that included concealed allocation assignment, adequate follow-up, and blinded outcome assessment. Treatments included operative (using a similar surgical protocol), casting for 6 weeks, and different types of functional intervention (bracing, wrapping, orthoses such as air casts, special shoes, and cast immobilization for 3 weeks). Three methods employing minimal treatment (less than 3 weeks of socks, braces, wraps, or casts) were also included because no trials of nonintervention were identified. Data were analyzed for homogeneity by calculating the Q statistic.
OUTCOMES MEASURED: The authors evaluated 2 long-term outcomes (residual pain and giving way) and one short-term outcome (time lost from work).
RESULTS: The natural history of untreated lateral ankle ligament ruptures is unknown. Treatments of short duration or inadequate ankle joint support are associated with more frequent residual symptoms (pain and giving way). Operative treatment followed by functional treatment resulted in less giving way than functional treatment alone (relative risk [RR]=0.23; 95% confidence interval [CI], 0.17-0.31). Functional treatment alone led to less giving way than 6-week cast treatment (RR=0.69; 95% CI, 0.50-0.94). Functional treatment was also associated with less pain than 6-week cast treatment (RR=0.67; 95% CI, 0.50-0.90). No other differences between interventions were found with regard to pain. Time lost from work varied greatly. Approximately 10% of patients undergoing operative treatment developed surgical complications. Results of a subgroup analysis of the high-quality studies only were not different from the results of the corresponding analysis of all studies.
To improve long-term ankle stability, patients with lateral ligament ruptures should be managed aggressively by either functional treatment or surgery followed by functional treatment. Because surgery is costly and associated with possible complications, a trial of functional treatment should be attempted first. If an initial nonsurgical approach fails, subsequent operative reconstruction is warranted.
BACKGROUND: The nearly 23,000 daily ankle injuries in the United States account for approximately 25% of all musculoskeletal injuries. The anterior talofibular ligament (a lateral supporting ligament) is most often involved. Current management options include casting, surgery, functional treatment, or minimal-to-no treatment. Despite recently published reviews suggesting that functional treatment and early mobilization provide the best results, other studies recommend operative treatment. This meta-analysis was designed to systematically determine the best treatment approach.
POPULATION STUDIED: More than 3300 patients with acute lateral ankle ligament ruptures enrolled in 27 randomized controlled trials were evaluated. Fifteen of the 42 trials initially identified were excluded because of poor follow-up (<40% dropout), republication of the same trial, or lack of relevant outcome variables. Diagnosis based on imaging by an arthrogram or stress radiograph was required. Study duration ranged from 6 to 46 months. Articles were included regardless of published language. Patients were largely recruited from the general population, so the results likely apply to those managed by family physicians.
STUDY DESIGN AND VALIDITY: The authors of this meta-analysis reviewed randomized controlled trials found by searching the MEDLINE, Cochrane, and EMBASE databases from 1966 to 1998. Three independent investigators blinded to author and institution reviewed each trial and assessed its quality. Quality was determined by randomization that included concealed allocation assignment, adequate follow-up, and blinded outcome assessment. Treatments included operative (using a similar surgical protocol), casting for 6 weeks, and different types of functional intervention (bracing, wrapping, orthoses such as air casts, special shoes, and cast immobilization for 3 weeks). Three methods employing minimal treatment (less than 3 weeks of socks, braces, wraps, or casts) were also included because no trials of nonintervention were identified. Data were analyzed for homogeneity by calculating the Q statistic.
OUTCOMES MEASURED: The authors evaluated 2 long-term outcomes (residual pain and giving way) and one short-term outcome (time lost from work).
RESULTS: The natural history of untreated lateral ankle ligament ruptures is unknown. Treatments of short duration or inadequate ankle joint support are associated with more frequent residual symptoms (pain and giving way). Operative treatment followed by functional treatment resulted in less giving way than functional treatment alone (relative risk [RR]=0.23; 95% confidence interval [CI], 0.17-0.31). Functional treatment alone led to less giving way than 6-week cast treatment (RR=0.69; 95% CI, 0.50-0.94). Functional treatment was also associated with less pain than 6-week cast treatment (RR=0.67; 95% CI, 0.50-0.90). No other differences between interventions were found with regard to pain. Time lost from work varied greatly. Approximately 10% of patients undergoing operative treatment developed surgical complications. Results of a subgroup analysis of the high-quality studies only were not different from the results of the corresponding analysis of all studies.
To improve long-term ankle stability, patients with lateral ligament ruptures should be managed aggressively by either functional treatment or surgery followed by functional treatment. Because surgery is costly and associated with possible complications, a trial of functional treatment should be attempted first. If an initial nonsurgical approach fails, subsequent operative reconstruction is warranted.
Which patients with blunt trauma do not require cervical spine x-rays?
BACKGROUND: Fear of occult cervical spine injury in patients who have experienced blunt trauma compels clinicians to liberally order cervical spine radiographs. Consequently, a high percentage of normal radiographs are obtained at a high monetary cost. The National Emergency X-Radiography Utilization Study (NEXUS) was conducted to validate a simple clinical prediction guide used to identify blunt trauma patients at low risk for cervical spine injury.
POPULATION STUDIED: All blunt trauma patients who had cervical x-rays in 21 emergency departments across the country were enrolled in this investigation. The study sites were located in hospitals that varied in size, level of activity of the emergency department, level of care, and other factors. The patients ranged in age from 1 year to 101 years; the mean age was 37 years.
STUDY DESIGN AND VALIDITY: This was a prospective validation of a clinical prediction guide. This guide rules out the need for cervical radiography if 5 clinical criteria are fulfilled: (1) absence of tenderness at the posterior midline of the cervical spine, (2) absence of a focal neurologic deficit, (3) a normal level of alertness, (4) no evidence of intoxication, and (5) absence of clinically apparent pain that might distract the patient from the pain of a cervical spine injury. Clinicians treating blunt trauma patients prospectively completed a study data form that included assessment of the 5 clinical criteria and demographic information. Each patient received cervical x-rays at the discretion of the treating physician, who was instructed to use his usual criteria for obtaining radiographs. Designated radiologists at each study site formally interpreted all radiographs while being blinded to the clinical assessment of the treating physician. The statistical analysis included basic performance measures of the decision instrument (sensitivity, specificity, positive predictive value, and negative predictive value). Two limitations of the study are that the 5 clinical criteria are open to some individual interpretation, and that “blunt trauma” was not defined. The treating clinicians were also not blinded to the decision instrument being tested.
OUTCOME MEASURED: The primary outcomes were the sensitivity, specificity, positive predictive value, and negative predictive value of the clinical prediction guide for the detection of radiographically confirmed cervical spine injury.
RESULTS: of 34,069 patients who underwent cervical spine radiography, 818 (2.4%) had documented cervical spine injuries. Although the clinical prediction guide failed to identify 8 of these injuries, only 2 were classified as clinically significant. The negative predictive value of the 5 clinical criteria for patients who did not have clinically significant injuries was 99.9% (95% confidence interval, 99.8%-100%). In other words, blunt trauma patients who screen negative for all 5 clinical criteria have a 99.9% chance of not having a clinically significant injury. The decision instrument identified 4306 patients (12.6%) as having low probability of cervical spine injury. These patients could have been spared radiographic evaluation.
The NEXUS study successfully validates a clinical prediction guide that identifies blunt trauma patients at low risk for cervical spine injury.
Blunt trauma patients who satisfy the 5 clinical criteria do not need cervical spine radiographs.
BACKGROUND: Fear of occult cervical spine injury in patients who have experienced blunt trauma compels clinicians to liberally order cervical spine radiographs. Consequently, a high percentage of normal radiographs are obtained at a high monetary cost. The National Emergency X-Radiography Utilization Study (NEXUS) was conducted to validate a simple clinical prediction guide used to identify blunt trauma patients at low risk for cervical spine injury.
POPULATION STUDIED: All blunt trauma patients who had cervical x-rays in 21 emergency departments across the country were enrolled in this investigation. The study sites were located in hospitals that varied in size, level of activity of the emergency department, level of care, and other factors. The patients ranged in age from 1 year to 101 years; the mean age was 37 years.
STUDY DESIGN AND VALIDITY: This was a prospective validation of a clinical prediction guide. This guide rules out the need for cervical radiography if 5 clinical criteria are fulfilled: (1) absence of tenderness at the posterior midline of the cervical spine, (2) absence of a focal neurologic deficit, (3) a normal level of alertness, (4) no evidence of intoxication, and (5) absence of clinically apparent pain that might distract the patient from the pain of a cervical spine injury. Clinicians treating blunt trauma patients prospectively completed a study data form that included assessment of the 5 clinical criteria and demographic information. Each patient received cervical x-rays at the discretion of the treating physician, who was instructed to use his usual criteria for obtaining radiographs. Designated radiologists at each study site formally interpreted all radiographs while being blinded to the clinical assessment of the treating physician. The statistical analysis included basic performance measures of the decision instrument (sensitivity, specificity, positive predictive value, and negative predictive value). Two limitations of the study are that the 5 clinical criteria are open to some individual interpretation, and that “blunt trauma” was not defined. The treating clinicians were also not blinded to the decision instrument being tested.
OUTCOME MEASURED: The primary outcomes were the sensitivity, specificity, positive predictive value, and negative predictive value of the clinical prediction guide for the detection of radiographically confirmed cervical spine injury.
RESULTS: of 34,069 patients who underwent cervical spine radiography, 818 (2.4%) had documented cervical spine injuries. Although the clinical prediction guide failed to identify 8 of these injuries, only 2 were classified as clinically significant. The negative predictive value of the 5 clinical criteria for patients who did not have clinically significant injuries was 99.9% (95% confidence interval, 99.8%-100%). In other words, blunt trauma patients who screen negative for all 5 clinical criteria have a 99.9% chance of not having a clinically significant injury. The decision instrument identified 4306 patients (12.6%) as having low probability of cervical spine injury. These patients could have been spared radiographic evaluation.
The NEXUS study successfully validates a clinical prediction guide that identifies blunt trauma patients at low risk for cervical spine injury.
Blunt trauma patients who satisfy the 5 clinical criteria do not need cervical spine radiographs.
BACKGROUND: Fear of occult cervical spine injury in patients who have experienced blunt trauma compels clinicians to liberally order cervical spine radiographs. Consequently, a high percentage of normal radiographs are obtained at a high monetary cost. The National Emergency X-Radiography Utilization Study (NEXUS) was conducted to validate a simple clinical prediction guide used to identify blunt trauma patients at low risk for cervical spine injury.
POPULATION STUDIED: All blunt trauma patients who had cervical x-rays in 21 emergency departments across the country were enrolled in this investigation. The study sites were located in hospitals that varied in size, level of activity of the emergency department, level of care, and other factors. The patients ranged in age from 1 year to 101 years; the mean age was 37 years.
STUDY DESIGN AND VALIDITY: This was a prospective validation of a clinical prediction guide. This guide rules out the need for cervical radiography if 5 clinical criteria are fulfilled: (1) absence of tenderness at the posterior midline of the cervical spine, (2) absence of a focal neurologic deficit, (3) a normal level of alertness, (4) no evidence of intoxication, and (5) absence of clinically apparent pain that might distract the patient from the pain of a cervical spine injury. Clinicians treating blunt trauma patients prospectively completed a study data form that included assessment of the 5 clinical criteria and demographic information. Each patient received cervical x-rays at the discretion of the treating physician, who was instructed to use his usual criteria for obtaining radiographs. Designated radiologists at each study site formally interpreted all radiographs while being blinded to the clinical assessment of the treating physician. The statistical analysis included basic performance measures of the decision instrument (sensitivity, specificity, positive predictive value, and negative predictive value). Two limitations of the study are that the 5 clinical criteria are open to some individual interpretation, and that “blunt trauma” was not defined. The treating clinicians were also not blinded to the decision instrument being tested.
OUTCOME MEASURED: The primary outcomes were the sensitivity, specificity, positive predictive value, and negative predictive value of the clinical prediction guide for the detection of radiographically confirmed cervical spine injury.
RESULTS: of 34,069 patients who underwent cervical spine radiography, 818 (2.4%) had documented cervical spine injuries. Although the clinical prediction guide failed to identify 8 of these injuries, only 2 were classified as clinically significant. The negative predictive value of the 5 clinical criteria for patients who did not have clinically significant injuries was 99.9% (95% confidence interval, 99.8%-100%). In other words, blunt trauma patients who screen negative for all 5 clinical criteria have a 99.9% chance of not having a clinically significant injury. The decision instrument identified 4306 patients (12.6%) as having low probability of cervical spine injury. These patients could have been spared radiographic evaluation.
The NEXUS study successfully validates a clinical prediction guide that identifies blunt trauma patients at low risk for cervical spine injury.
Blunt trauma patients who satisfy the 5 clinical criteria do not need cervical spine radiographs.
Is azithromycin monotherapy as efficacious as cefuroxime plus erythromycin for the treatment of community-acquired pneumonia in hospitalized patients?
BACKGROUND: The American Thoracic Society recommends a second- or third-generation cephalosporin with or without a macrolide for empiric treatment of community-acquired pneumonia (CAP) in hospitalized patients. These guidelines have resulted in better outcomes in patients younger than 60 years but not in older patients.1 Azithromycin has demonstrated in vitro and in vivo activity against the common pathogens in CAP, including Streptococcus pneumoniae, Haemophilus influenzae, Legionella pneumophila, and Mycoplasma pneumoniae. This once-a-day macrolide possesses good tissue penetration and sustained tissue levels that allow for a shorter course of therapy.
POPULATION STUDIED: This multicenter study reported the results of 145 evaluable hospitalized patients with a primary diagnosis of CAP. The average age of these patients was not stated. Their condition was diagnosed by chest x-ray, with either one or more signs and symptoms consistent with lower respiratory infection or an elevated white blood cell count. Patients were excluded if they were allergic to the study drugs, had been hospitalized during the previous 14 days, were immunosuppressed, had serious renal dysfunction or conditions affecting drug absorption, or took an antibiotic within 24 hours of enrollment.
STUDY DESIGN AND VALIDITY: This was a prospective randomized unblinded trial comparing the empiric treatment of CAP with 2 regimens. One group of patients received intravenous azithromycin 500 mg daily for 2 to 5 days followed by oral therapy to complete a 7- to 10-day course. The second group was treated with the combination of intravenous cefuroxime 750 mg every 8 hours for 2 to 7 days followed by oral cefuroxime 500 mg twice a day for a complete 7- to 10-day course along with erythromycin lactobionate or base 500 mg to 1 g intravenously every 6 hours and then orally for a total of 21 days. This study was well done, with several limitations. Concealed allocation of randomization occurred through the use of opaque envelopes after each patient was stratified by risk factors. The sample size was large enough to find a 14% difference in clinical response rates with a power of 80%. Analysis was by intention to treat. Few data (such as all-important age) were provided about the type of patients in this study, making it difficult to make generalizations. Also, the evaluators of clinical response were aware of what treatment the patient received. Given the subjective nature of the end points of “cure” versus “failure,” this lack of blinding could have affected these outcome measures. However, end points such as “time to cure” (not evaluated in this study) would be more likely to be affected by observer bias.
OUTCOMES MEASURED: Clinical response (success/failure) was the primary outcome. Adverse events and in-hospital mortality were secondary end points.
RESULTS: The causative infecting organism was not known in 41% of the patients. The most common causative agents in patients with definitive or presumptive diagnoses were: S pneumonaie (19%), L pneumophila (14%), and H influenzae (13%). The rate of cure of patients with CAP was similar: azithromycin 75% (95% confidence interval [CI], 64%-84%) and cefuroxime plus erythromycin 83% (95% CI, 73%-90%). Cure rates did not differ by causative organism. Adverse effects occurred significantly more often in the cefuroxime/erythromycin-treated patients (P <.001), due in large part to catheter site reactions and gastrointestinal tract symptoms presumably related to erythromycin. Mortality was similar in the 2 groups
For the treatment of inpatient CAP, azithromycin is just as effective as the combination of cefuroxime plus erythromycin. Its favorable side effect profile and cost makes azithromycin a reasonable choice for the treatment of CAP in hospitalized patients.
BACKGROUND: The American Thoracic Society recommends a second- or third-generation cephalosporin with or without a macrolide for empiric treatment of community-acquired pneumonia (CAP) in hospitalized patients. These guidelines have resulted in better outcomes in patients younger than 60 years but not in older patients.1 Azithromycin has demonstrated in vitro and in vivo activity against the common pathogens in CAP, including Streptococcus pneumoniae, Haemophilus influenzae, Legionella pneumophila, and Mycoplasma pneumoniae. This once-a-day macrolide possesses good tissue penetration and sustained tissue levels that allow for a shorter course of therapy.
POPULATION STUDIED: This multicenter study reported the results of 145 evaluable hospitalized patients with a primary diagnosis of CAP. The average age of these patients was not stated. Their condition was diagnosed by chest x-ray, with either one or more signs and symptoms consistent with lower respiratory infection or an elevated white blood cell count. Patients were excluded if they were allergic to the study drugs, had been hospitalized during the previous 14 days, were immunosuppressed, had serious renal dysfunction or conditions affecting drug absorption, or took an antibiotic within 24 hours of enrollment.
STUDY DESIGN AND VALIDITY: This was a prospective randomized unblinded trial comparing the empiric treatment of CAP with 2 regimens. One group of patients received intravenous azithromycin 500 mg daily for 2 to 5 days followed by oral therapy to complete a 7- to 10-day course. The second group was treated with the combination of intravenous cefuroxime 750 mg every 8 hours for 2 to 7 days followed by oral cefuroxime 500 mg twice a day for a complete 7- to 10-day course along with erythromycin lactobionate or base 500 mg to 1 g intravenously every 6 hours and then orally for a total of 21 days. This study was well done, with several limitations. Concealed allocation of randomization occurred through the use of opaque envelopes after each patient was stratified by risk factors. The sample size was large enough to find a 14% difference in clinical response rates with a power of 80%. Analysis was by intention to treat. Few data (such as all-important age) were provided about the type of patients in this study, making it difficult to make generalizations. Also, the evaluators of clinical response were aware of what treatment the patient received. Given the subjective nature of the end points of “cure” versus “failure,” this lack of blinding could have affected these outcome measures. However, end points such as “time to cure” (not evaluated in this study) would be more likely to be affected by observer bias.
OUTCOMES MEASURED: Clinical response (success/failure) was the primary outcome. Adverse events and in-hospital mortality were secondary end points.
RESULTS: The causative infecting organism was not known in 41% of the patients. The most common causative agents in patients with definitive or presumptive diagnoses were: S pneumonaie (19%), L pneumophila (14%), and H influenzae (13%). The rate of cure of patients with CAP was similar: azithromycin 75% (95% confidence interval [CI], 64%-84%) and cefuroxime plus erythromycin 83% (95% CI, 73%-90%). Cure rates did not differ by causative organism. Adverse effects occurred significantly more often in the cefuroxime/erythromycin-treated patients (P <.001), due in large part to catheter site reactions and gastrointestinal tract symptoms presumably related to erythromycin. Mortality was similar in the 2 groups
For the treatment of inpatient CAP, azithromycin is just as effective as the combination of cefuroxime plus erythromycin. Its favorable side effect profile and cost makes azithromycin a reasonable choice for the treatment of CAP in hospitalized patients.
BACKGROUND: The American Thoracic Society recommends a second- or third-generation cephalosporin with or without a macrolide for empiric treatment of community-acquired pneumonia (CAP) in hospitalized patients. These guidelines have resulted in better outcomes in patients younger than 60 years but not in older patients.1 Azithromycin has demonstrated in vitro and in vivo activity against the common pathogens in CAP, including Streptococcus pneumoniae, Haemophilus influenzae, Legionella pneumophila, and Mycoplasma pneumoniae. This once-a-day macrolide possesses good tissue penetration and sustained tissue levels that allow for a shorter course of therapy.
POPULATION STUDIED: This multicenter study reported the results of 145 evaluable hospitalized patients with a primary diagnosis of CAP. The average age of these patients was not stated. Their condition was diagnosed by chest x-ray, with either one or more signs and symptoms consistent with lower respiratory infection or an elevated white blood cell count. Patients were excluded if they were allergic to the study drugs, had been hospitalized during the previous 14 days, were immunosuppressed, had serious renal dysfunction or conditions affecting drug absorption, or took an antibiotic within 24 hours of enrollment.
STUDY DESIGN AND VALIDITY: This was a prospective randomized unblinded trial comparing the empiric treatment of CAP with 2 regimens. One group of patients received intravenous azithromycin 500 mg daily for 2 to 5 days followed by oral therapy to complete a 7- to 10-day course. The second group was treated with the combination of intravenous cefuroxime 750 mg every 8 hours for 2 to 7 days followed by oral cefuroxime 500 mg twice a day for a complete 7- to 10-day course along with erythromycin lactobionate or base 500 mg to 1 g intravenously every 6 hours and then orally for a total of 21 days. This study was well done, with several limitations. Concealed allocation of randomization occurred through the use of opaque envelopes after each patient was stratified by risk factors. The sample size was large enough to find a 14% difference in clinical response rates with a power of 80%. Analysis was by intention to treat. Few data (such as all-important age) were provided about the type of patients in this study, making it difficult to make generalizations. Also, the evaluators of clinical response were aware of what treatment the patient received. Given the subjective nature of the end points of “cure” versus “failure,” this lack of blinding could have affected these outcome measures. However, end points such as “time to cure” (not evaluated in this study) would be more likely to be affected by observer bias.
OUTCOMES MEASURED: Clinical response (success/failure) was the primary outcome. Adverse events and in-hospital mortality were secondary end points.
RESULTS: The causative infecting organism was not known in 41% of the patients. The most common causative agents in patients with definitive or presumptive diagnoses were: S pneumonaie (19%), L pneumophila (14%), and H influenzae (13%). The rate of cure of patients with CAP was similar: azithromycin 75% (95% confidence interval [CI], 64%-84%) and cefuroxime plus erythromycin 83% (95% CI, 73%-90%). Cure rates did not differ by causative organism. Adverse effects occurred significantly more often in the cefuroxime/erythromycin-treated patients (P <.001), due in large part to catheter site reactions and gastrointestinal tract symptoms presumably related to erythromycin. Mortality was similar in the 2 groups
For the treatment of inpatient CAP, azithromycin is just as effective as the combination of cefuroxime plus erythromycin. Its favorable side effect profile and cost makes azithromycin a reasonable choice for the treatment of CAP in hospitalized patients.
Is exposure to HMG-CoA reductase inhibitors (statins), fibrates, or other lipid-lowering drugs associated with reduced risk of bone fracture in older patients?
BACKGROUND: Animal studies have demonstrated that 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) increase bone formation and bone volume. Although studies in humans suggest that statins may increase bone formation, no research has been done to determine whether older persons taking these drugs will have a decreased risk of fracture.
POPULATION STUDIED: The study used data from the United Kingdom-based General Practice Research Database, which includes information from approximately 300 general practices. The base population consisted of more than 91,000 individuals aged 50 to 89 years. Of these, more than 28,000 had received at least 1 prescription for a lipid-lowering drug for hyperlipidemia, approximately 13,000 had hyperlipidemia but were not taking lipid-lowering agents, and 50,000 were randomly selected patients without a diagnosis of hyperlipidemia. Patients were excluded if they had a diagnosis of osteoporosis, osteomalacia, malignancy, alcoholism, or if they were taking bisphosphonates. A total of 3940 patients who had a bone fracture were identified by the International Classification of Diseases-eighth revision (ICD-8) and were verified by hospital discharge records or referral letters to specialists. These case patients were matched for age, sex, practice attended, calendar year, and years enrolled to 23,379 control group patients from the base population who had not had a fracture.
STUDY DESIGN AND VALIDITY: This was a large population-based case-control study nested within an ongoing study. Members of the 3 groups were followed until they developed a fracture, left the practice, or died. Patients taking a lipid-lowering drug were categorized as being current users, recent users, and past users.
OUTCOMES MEASURED: The primary outcome was the risk of development of a first-time fracture of the femur, humerus, hand, wrist, lower arm, clavicle, vertebral body, foot, or other unspecified fracture among the 3 groups as estimated by the odds ratio. The authors used statistical analysis (logistic regression) to consider the type, timing, and duration of treatment as estimated by numbers of prescriptions filled.
RESULTS: After adjusting for variables including smoking status, weight, hormone replacement therapy or corticosteroid use, and the number of office visits before the fracture, patients who had any current or recent recorded prescriptions for statins before the date of fracture had a significantly lowered risk of fracture with adjusted odds ratios of 0.55 for current users (95% confidence interval [CI], 0.44-0.69) and 0.67 (95% CI, 0.50-0.92) for recent users of statins. The odds ratio for previous users of statins was not statistically significant. A consistent class effect was noted, with all statins associated with decreased fracture risk. Other lipid-lowering agents were not associated with a reduced risk for fracture. The effect of statins on risk of fracture did not differ by the skeletal site of fracture, age group, or sex.
This population-based case-control study suggests that the HMG-CoA reductase inhibitors, but not other lipid-lowering agents, are associated with a significantly decreased risk of subsequent fracture in humans, especially in current users. There may be biologic plausibility for this effect in that both statins and bisphosphonates act in the same metabolic pathway (the mevalonate pathway) that may ultimately affect osteoclastic activity. Other recently reported case-controlled studies find protective effects of similar magnitude. Statin use has been associated with a 43% to 50% lowering of risk of hip fracture among elderly users, and another study reports a reduction of risk of fracture of 52% among elderly women using statins. 1,2
However, these studies cannot establish a causal relationship between statin use and decreased risk of fracture. As with all observational studies, it is possible that unknown factors may provide alternative explanations for these findings. Large randomized controlled trials are needed to confirm any benefit of statins in reducing risk of fracture. In addition, the added benefit of statins to current regimens for fracture prevention (calcium, Vitamin D, estrogen, biphosphonat
BACKGROUND: Animal studies have demonstrated that 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) increase bone formation and bone volume. Although studies in humans suggest that statins may increase bone formation, no research has been done to determine whether older persons taking these drugs will have a decreased risk of fracture.
POPULATION STUDIED: The study used data from the United Kingdom-based General Practice Research Database, which includes information from approximately 300 general practices. The base population consisted of more than 91,000 individuals aged 50 to 89 years. Of these, more than 28,000 had received at least 1 prescription for a lipid-lowering drug for hyperlipidemia, approximately 13,000 had hyperlipidemia but were not taking lipid-lowering agents, and 50,000 were randomly selected patients without a diagnosis of hyperlipidemia. Patients were excluded if they had a diagnosis of osteoporosis, osteomalacia, malignancy, alcoholism, or if they were taking bisphosphonates. A total of 3940 patients who had a bone fracture were identified by the International Classification of Diseases-eighth revision (ICD-8) and were verified by hospital discharge records or referral letters to specialists. These case patients were matched for age, sex, practice attended, calendar year, and years enrolled to 23,379 control group patients from the base population who had not had a fracture.
STUDY DESIGN AND VALIDITY: This was a large population-based case-control study nested within an ongoing study. Members of the 3 groups were followed until they developed a fracture, left the practice, or died. Patients taking a lipid-lowering drug were categorized as being current users, recent users, and past users.
OUTCOMES MEASURED: The primary outcome was the risk of development of a first-time fracture of the femur, humerus, hand, wrist, lower arm, clavicle, vertebral body, foot, or other unspecified fracture among the 3 groups as estimated by the odds ratio. The authors used statistical analysis (logistic regression) to consider the type, timing, and duration of treatment as estimated by numbers of prescriptions filled.
RESULTS: After adjusting for variables including smoking status, weight, hormone replacement therapy or corticosteroid use, and the number of office visits before the fracture, patients who had any current or recent recorded prescriptions for statins before the date of fracture had a significantly lowered risk of fracture with adjusted odds ratios of 0.55 for current users (95% confidence interval [CI], 0.44-0.69) and 0.67 (95% CI, 0.50-0.92) for recent users of statins. The odds ratio for previous users of statins was not statistically significant. A consistent class effect was noted, with all statins associated with decreased fracture risk. Other lipid-lowering agents were not associated with a reduced risk for fracture. The effect of statins on risk of fracture did not differ by the skeletal site of fracture, age group, or sex.
This population-based case-control study suggests that the HMG-CoA reductase inhibitors, but not other lipid-lowering agents, are associated with a significantly decreased risk of subsequent fracture in humans, especially in current users. There may be biologic plausibility for this effect in that both statins and bisphosphonates act in the same metabolic pathway (the mevalonate pathway) that may ultimately affect osteoclastic activity. Other recently reported case-controlled studies find protective effects of similar magnitude. Statin use has been associated with a 43% to 50% lowering of risk of hip fracture among elderly users, and another study reports a reduction of risk of fracture of 52% among elderly women using statins. 1,2
However, these studies cannot establish a causal relationship between statin use and decreased risk of fracture. As with all observational studies, it is possible that unknown factors may provide alternative explanations for these findings. Large randomized controlled trials are needed to confirm any benefit of statins in reducing risk of fracture. In addition, the added benefit of statins to current regimens for fracture prevention (calcium, Vitamin D, estrogen, biphosphonat
BACKGROUND: Animal studies have demonstrated that 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) increase bone formation and bone volume. Although studies in humans suggest that statins may increase bone formation, no research has been done to determine whether older persons taking these drugs will have a decreased risk of fracture.
POPULATION STUDIED: The study used data from the United Kingdom-based General Practice Research Database, which includes information from approximately 300 general practices. The base population consisted of more than 91,000 individuals aged 50 to 89 years. Of these, more than 28,000 had received at least 1 prescription for a lipid-lowering drug for hyperlipidemia, approximately 13,000 had hyperlipidemia but were not taking lipid-lowering agents, and 50,000 were randomly selected patients without a diagnosis of hyperlipidemia. Patients were excluded if they had a diagnosis of osteoporosis, osteomalacia, malignancy, alcoholism, or if they were taking bisphosphonates. A total of 3940 patients who had a bone fracture were identified by the International Classification of Diseases-eighth revision (ICD-8) and were verified by hospital discharge records or referral letters to specialists. These case patients were matched for age, sex, practice attended, calendar year, and years enrolled to 23,379 control group patients from the base population who had not had a fracture.
STUDY DESIGN AND VALIDITY: This was a large population-based case-control study nested within an ongoing study. Members of the 3 groups were followed until they developed a fracture, left the practice, or died. Patients taking a lipid-lowering drug were categorized as being current users, recent users, and past users.
OUTCOMES MEASURED: The primary outcome was the risk of development of a first-time fracture of the femur, humerus, hand, wrist, lower arm, clavicle, vertebral body, foot, or other unspecified fracture among the 3 groups as estimated by the odds ratio. The authors used statistical analysis (logistic regression) to consider the type, timing, and duration of treatment as estimated by numbers of prescriptions filled.
RESULTS: After adjusting for variables including smoking status, weight, hormone replacement therapy or corticosteroid use, and the number of office visits before the fracture, patients who had any current or recent recorded prescriptions for statins before the date of fracture had a significantly lowered risk of fracture with adjusted odds ratios of 0.55 for current users (95% confidence interval [CI], 0.44-0.69) and 0.67 (95% CI, 0.50-0.92) for recent users of statins. The odds ratio for previous users of statins was not statistically significant. A consistent class effect was noted, with all statins associated with decreased fracture risk. Other lipid-lowering agents were not associated with a reduced risk for fracture. The effect of statins on risk of fracture did not differ by the skeletal site of fracture, age group, or sex.
This population-based case-control study suggests that the HMG-CoA reductase inhibitors, but not other lipid-lowering agents, are associated with a significantly decreased risk of subsequent fracture in humans, especially in current users. There may be biologic plausibility for this effect in that both statins and bisphosphonates act in the same metabolic pathway (the mevalonate pathway) that may ultimately affect osteoclastic activity. Other recently reported case-controlled studies find protective effects of similar magnitude. Statin use has been associated with a 43% to 50% lowering of risk of hip fracture among elderly users, and another study reports a reduction of risk of fracture of 52% among elderly women using statins. 1,2
However, these studies cannot establish a causal relationship between statin use and decreased risk of fracture. As with all observational studies, it is possible that unknown factors may provide alternative explanations for these findings. Large randomized controlled trials are needed to confirm any benefit of statins in reducing risk of fracture. In addition, the added benefit of statins to current regimens for fracture prevention (calcium, Vitamin D, estrogen, biphosphonat