Warren Newton, MD, MPH

ABSTRACT

BACKGROUND: Many new oral medications have been developed to treat diabetes, but uncertainty remains regarding which are best for initial treatment and whether effectiveness rates differ. This review compares the available oral antihyperglycemics.

POPULATION STUDIED: A total of 63 randomized controlled clinical trials involving oral hypoglycemic drugs for type 2 diabetes was identified by a MED-LINE search and review of the bibliographies of articles found initially. Other inclusion criteria were study duration of at least 3 months, at least 10 subjects at the study’s conclusion, and hemoglobin A_1C levels reported. Other search details, such as the year and key words of a study, were not mentioned. More than 15,000 subjects have been enrolled in the identified trials, but no information was given regarding important clinical characteristics such as age, ethnicity, body mass index, or medical conditions other than diabetes. Therefore, assessing generalizability of the data to typical patients of family practitioners is difficult.

STUDY DESIGN AND VALIDITY: The article lists available randomized clinical trials that evaluate sulfonylureas, metformin, α-glucosidase inhibitors (AGIs), thiazolidinediones (TZDs), and nonsulfonylurea secretagogues as monotherapy versus placebo, in head-to-head trials or in combination, and compares their outcomes in terms of hemoglobin A_1C reduction. When multiple doses of a drug were tested, the results from the highest dose were used. There was no attempt to synthesize the data provided by the studies into a meta-analysis.

OUTCOMES MEASURED: The major outcome measured was percent hemoglobin A_1C reduction. Side effects were mentioned but not quantified. Cost, patient satisfaction, and quality of life were not addressed.

RESULTS: Except for the UKPDS, all available studies of oral hypoglycemics are short term and are limited in focus to hemoglobin A_1C. Each class of drugs achieved a similar initial reduction in hemoglobin A_1C of 1% to 2% except for the AGIs and nateglinide, which were less effective. The results are remarkably consistent across studies. Head-to-head comparison of specific medications further supports this conclusion. When taken in combination, the effects on hemoglobin A_1C are additive.

ABSTRACT

BACKGROUND: Many new oral medications have been developed to treat diabetes, but uncertainty remains regarding which are best for initial treatment and whether effectiveness rates differ. This review compares the available oral antihyperglycemics.

POPULATION STUDIED: A total of 63 randomized controlled clinical trials involving oral hypoglycemic drugs for type 2 diabetes was identified by a MED-LINE search and review of the bibliographies of articles found initially. Other inclusion criteria were study duration of at least 3 months, at least 10 subjects at the study’s conclusion, and hemoglobin A_1C levels reported. Other search details, such as the year and key words of a study, were not mentioned. More than 15,000 subjects have been enrolled in the identified trials, but no information was given regarding important clinical characteristics such as age, ethnicity, body mass index, or medical conditions other than diabetes. Therefore, assessing generalizability of the data to typical patients of family practitioners is difficult.

STUDY DESIGN AND VALIDITY: The article lists available randomized clinical trials that evaluate sulfonylureas, metformin, α-glucosidase inhibitors (AGIs), thiazolidinediones (TZDs), and nonsulfonylurea secretagogues as monotherapy versus placebo, in head-to-head trials or in combination, and compares their outcomes in terms of hemoglobin A_1C reduction. When multiple doses of a drug were tested, the results from the highest dose were used. There was no attempt to synthesize the data provided by the studies into a meta-analysis.

OUTCOMES MEASURED: The major outcome measured was percent hemoglobin A_1C reduction. Side effects were mentioned but not quantified. Cost, patient satisfaction, and quality of life were not addressed.

RESULTS: Except for the UKPDS, all available studies of oral hypoglycemics are short term and are limited in focus to hemoglobin A_1C. Each class of drugs achieved a similar initial reduction in hemoglobin A_1C of 1% to 2% except for the AGIs and nateglinide, which were less effective. The results are remarkably consistent across studies. Head-to-head comparison of specific medications further supports this conclusion. When taken in combination, the effects on hemoglobin A_1C are additive.

ABSTRACT

BACKGROUND: Many new oral medications have been developed to treat diabetes, but uncertainty remains regarding which are best for initial treatment and whether effectiveness rates differ. This review compares the available oral antihyperglycemics.

POPULATION STUDIED: A total of 63 randomized controlled clinical trials involving oral hypoglycemic drugs for type 2 diabetes was identified by a MED-LINE search and review of the bibliographies of articles found initially. Other inclusion criteria were study duration of at least 3 months, at least 10 subjects at the study’s conclusion, and hemoglobin A_1C levels reported. Other search details, such as the year and key words of a study, were not mentioned. More than 15,000 subjects have been enrolled in the identified trials, but no information was given regarding important clinical characteristics such as age, ethnicity, body mass index, or medical conditions other than diabetes. Therefore, assessing generalizability of the data to typical patients of family practitioners is difficult.

STUDY DESIGN AND VALIDITY: The article lists available randomized clinical trials that evaluate sulfonylureas, metformin, α-glucosidase inhibitors (AGIs), thiazolidinediones (TZDs), and nonsulfonylurea secretagogues as monotherapy versus placebo, in head-to-head trials or in combination, and compares their outcomes in terms of hemoglobin A_1C reduction. When multiple doses of a drug were tested, the results from the highest dose were used. There was no attempt to synthesize the data provided by the studies into a meta-analysis.

OUTCOMES MEASURED: The major outcome measured was percent hemoglobin A_1C reduction. Side effects were mentioned but not quantified. Cost, patient satisfaction, and quality of life were not addressed.

RESULTS: Except for the UKPDS, all available studies of oral hypoglycemics are short term and are limited in focus to hemoglobin A_1C. Each class of drugs achieved a similar initial reduction in hemoglobin A_1C of 1% to 2% except for the AGIs and nateglinide, which were less effective. The results are remarkably consistent across studies. Head-to-head comparison of specific medications further supports this conclusion. When taken in combination, the effects on hemoglobin A_1C are additive.

BACKGROUND: Long-acting anti-angina agents have theoretical advantages over short-acting options. They offer more consistent blood levels, which could result in fewer side effects, as well as more sustained control of symptoms. This randomized controlled trial addressed the effectiveness of a strategy of converting to long-acting antianginal medication.

POPULATION STUDIED: The investigators enrolled 100 male outpatients at a Veterans Administration (VA) Health Systems Clinic who had known coronary disease or angina and who were taking at least 2 antianginal drugs. They were identified by clinician or a pharmacy database. Patients were excluded from the study if they had a recent hospitalization, aortic stenosis, an ejection fraction of less than 40%, significant conduction defects, or limited life expectancy. Most of the men (81%) were white and the average age was 65 years. The men had several concomitant diseases: 39% had diabetes, 28% congestive heart failure, and 68% had a prior myocardial infarction. These patients seem similar to the sicker patients in a typical family practice, but caution should be exercised in generalizing the results to women and to settings different from the VA.

STUDY DESIGN AND VALIDITY: The study was a single-blind, prospective randomized trial using concealed allocation. A nurse practitioner evaluated all patients weekly for 4 weeks and then monthly for 2 months. In the “once a day” group, all previous antianginal medications were withdrawn and then restarted one at a time in this order: long-acting diltiazem (up to 360 mg/day), nitroglycerin patches (up to 0.8 mg/hour) and atenolol (up to 100 mg/day). Doses were maximized prior to the addition of the next drug. In contrast, no drugs were stopped in “usual care”; instead, baseline medications were increased on the basis of symptoms to maximum doses following alphabetical order. For both groups, an algorithm was used to adjust medications. Data were analyzed according to intention-to-treat using t tests, with logistic regression to control for baseline frequency of angina. The methodology of this study was pragmatic and appropriate. Strengths included randomization, concealed allocation and complete follow-up. A major limitation of the design was that the intervention was complex, involving frequent visits along with a specific care algorithm, complete withdrawal of medication at the beginning and a specified order of medication re-introduction. The consequence is that it is difficult to determine which component of the intervention led to the results. Other important limitations include short duration (3 months); inattention to important confounding variables such as diabetes, congestive heart failure, or which specific drugs the patients were taking; lack of statistical power to assess confounding; and lack of attention to multiple comparisons.

OUTCOMES MEASURED: Primary outcomes were functional status, treatment satisfaction, and quality of life, measured by the Seattle Angina Questionnaire, a validated disease-specific measure for patients with coronary artery disease. The study did not address cost, side-effects, morbidity/mortality, or any long-term outcomes.

RESULTS: At the end of the trial, the once-a-day group had fewer symptoms than the usual care group (difference of averages: 12.3 points, P <.002; 5-8 points is considered clinically significant). Controlling for baseline frequency of angina did not change this result. There was no significant difference between groups in treatment satisfaction or quality of life. By the end of the trial, the patients in the once-a-day group were taking fewer medications (average 1.55 vs 2.14, P <.001).

BACKGROUND: Long-acting anti-angina agents have theoretical advantages over short-acting options. They offer more consistent blood levels, which could result in fewer side effects, as well as more sustained control of symptoms. This randomized controlled trial addressed the effectiveness of a strategy of converting to long-acting antianginal medication.

POPULATION STUDIED: The investigators enrolled 100 male outpatients at a Veterans Administration (VA) Health Systems Clinic who had known coronary disease or angina and who were taking at least 2 antianginal drugs. They were identified by clinician or a pharmacy database. Patients were excluded from the study if they had a recent hospitalization, aortic stenosis, an ejection fraction of less than 40%, significant conduction defects, or limited life expectancy. Most of the men (81%) were white and the average age was 65 years. The men had several concomitant diseases: 39% had diabetes, 28% congestive heart failure, and 68% had a prior myocardial infarction. These patients seem similar to the sicker patients in a typical family practice, but caution should be exercised in generalizing the results to women and to settings different from the VA.

STUDY DESIGN AND VALIDITY: The study was a single-blind, prospective randomized trial using concealed allocation. A nurse practitioner evaluated all patients weekly for 4 weeks and then monthly for 2 months. In the “once a day” group, all previous antianginal medications were withdrawn and then restarted one at a time in this order: long-acting diltiazem (up to 360 mg/day), nitroglycerin patches (up to 0.8 mg/hour) and atenolol (up to 100 mg/day). Doses were maximized prior to the addition of the next drug. In contrast, no drugs were stopped in “usual care”; instead, baseline medications were increased on the basis of symptoms to maximum doses following alphabetical order. For both groups, an algorithm was used to adjust medications. Data were analyzed according to intention-to-treat using t tests, with logistic regression to control for baseline frequency of angina. The methodology of this study was pragmatic and appropriate. Strengths included randomization, concealed allocation and complete follow-up. A major limitation of the design was that the intervention was complex, involving frequent visits along with a specific care algorithm, complete withdrawal of medication at the beginning and a specified order of medication re-introduction. The consequence is that it is difficult to determine which component of the intervention led to the results. Other important limitations include short duration (3 months); inattention to important confounding variables such as diabetes, congestive heart failure, or which specific drugs the patients were taking; lack of statistical power to assess confounding; and lack of attention to multiple comparisons.

OUTCOMES MEASURED: Primary outcomes were functional status, treatment satisfaction, and quality of life, measured by the Seattle Angina Questionnaire, a validated disease-specific measure for patients with coronary artery disease. The study did not address cost, side-effects, morbidity/mortality, or any long-term outcomes.

RESULTS: At the end of the trial, the once-a-day group had fewer symptoms than the usual care group (difference of averages: 12.3 points, P <.002; 5-8 points is considered clinically significant). Controlling for baseline frequency of angina did not change this result. There was no significant difference between groups in treatment satisfaction or quality of life. By the end of the trial, the patients in the once-a-day group were taking fewer medications (average 1.55 vs 2.14, P <.001).

BACKGROUND: Long-acting anti-angina agents have theoretical advantages over short-acting options. They offer more consistent blood levels, which could result in fewer side effects, as well as more sustained control of symptoms. This randomized controlled trial addressed the effectiveness of a strategy of converting to long-acting antianginal medication.

POPULATION STUDIED: The investigators enrolled 100 male outpatients at a Veterans Administration (VA) Health Systems Clinic who had known coronary disease or angina and who were taking at least 2 antianginal drugs. They were identified by clinician or a pharmacy database. Patients were excluded from the study if they had a recent hospitalization, aortic stenosis, an ejection fraction of less than 40%, significant conduction defects, or limited life expectancy. Most of the men (81%) were white and the average age was 65 years. The men had several concomitant diseases: 39% had diabetes, 28% congestive heart failure, and 68% had a prior myocardial infarction. These patients seem similar to the sicker patients in a typical family practice, but caution should be exercised in generalizing the results to women and to settings different from the VA.

STUDY DESIGN AND VALIDITY: The study was a single-blind, prospective randomized trial using concealed allocation. A nurse practitioner evaluated all patients weekly for 4 weeks and then monthly for 2 months. In the “once a day” group, all previous antianginal medications were withdrawn and then restarted one at a time in this order: long-acting diltiazem (up to 360 mg/day), nitroglycerin patches (up to 0.8 mg/hour) and atenolol (up to 100 mg/day). Doses were maximized prior to the addition of the next drug. In contrast, no drugs were stopped in “usual care”; instead, baseline medications were increased on the basis of symptoms to maximum doses following alphabetical order. For both groups, an algorithm was used to adjust medications. Data were analyzed according to intention-to-treat using t tests, with logistic regression to control for baseline frequency of angina. The methodology of this study was pragmatic and appropriate. Strengths included randomization, concealed allocation and complete follow-up. A major limitation of the design was that the intervention was complex, involving frequent visits along with a specific care algorithm, complete withdrawal of medication at the beginning and a specified order of medication re-introduction. The consequence is that it is difficult to determine which component of the intervention led to the results. Other important limitations include short duration (3 months); inattention to important confounding variables such as diabetes, congestive heart failure, or which specific drugs the patients were taking; lack of statistical power to assess confounding; and lack of attention to multiple comparisons.

OUTCOMES MEASURED: Primary outcomes were functional status, treatment satisfaction, and quality of life, measured by the Seattle Angina Questionnaire, a validated disease-specific measure for patients with coronary artery disease. The study did not address cost, side-effects, morbidity/mortality, or any long-term outcomes.

RESULTS: At the end of the trial, the once-a-day group had fewer symptoms than the usual care group (difference of averages: 12.3 points, P <.002; 5-8 points is considered clinically significant). Controlling for baseline frequency of angina did not change this result. There was no significant difference between groups in treatment satisfaction or quality of life. By the end of the trial, the patients in the once-a-day group were taking fewer medications (average 1.55 vs 2.14, P <.001).

BACKGROUND: Pharmacologic methods of cervical ripening are associated with uterine contractile abnormalities. This randomized trial compares intravaginal misoprostol to transcervical Foley catheters for preinduction cervical ripening.

POPULATION STUDIED: A total of 111 pregnant women were admitted to a tertiary hospital for induction. Subjects had singleton vertex pregnancies at more than 28 weeks’ gestation and a Bishop score of less than 6; indications for induction included preeclampsia, oligohydramnios, postdates, and growth restriction. Exclusion criteria included previous uterine surgeries, rupture of membranes, previous induction, or use of a pre-induction agent during the pregnancy. Mean age was 26 years; 70% were nulliparous. Labor was managed actively with rupture of membranes as soon as possible, use of oxytocin, and an epidural rate of almost 90%. Although the study population included premature infants and the setting was different from that in which most family physicians deliver babies, the results probably apply to patients cared for by family physicians.

STUDY DESIGN AND VALIDITY: The participants were randomized with concealed allocation to placement of a transcervical Foley catheter or 50 mg intravaginal misoprostol. A 16F Foley catheter was inserted under direct visualization during a sterile speculum examination without an obturator, tenaculum, or cervical cleansing. Once past the internal os, the balloon was filled with 30 cc of sterile water, and gentle traction was applied (the end of the catheter was taped to the patient’s medial knee or thigh). Every 6 hours the catheter was adjusted to continue traction. In the misoprostol group, 50 μg was placed in the posterior fornix of the vagina every 4 hours up to 6 times. Oxytocin was begun after extrusion of the catheter or 4 hours after the last dose of misoprostol if no labor ensued or if all 6 doses had been given. Bishop scores were assigned before randomization and at the time of Foley catheter extrusion or last dose of misoprostol. The Student t test, Fisher exact test, Mann-Whitney U, and chi square were used to compare groups; no information was given about whether testing was by intention to treat. The methodologic strength of this study was adequate. Its major strengths were randomization with concealed allocation and adequate power. Major weaknesses included the lack of attention to potential confounding factors, such as labor support, that are known to influence the course of labor and the inability to mask the physician assessing the Bishop scores. Other minor weaknesses included the intrinsic lack of intra- and inter-observer reliability of the Bishop score, the possible variation in the timing of giving the Bishop score for the misoprostol group, and the lack of power for assessment of outcomes other than the primary ones.

OUTCOMES MEASURED: The primary outcome measure was change in Bishop score. Secondary outcome measures included length of time for preinduction cervical ripening, total time for induction, delivery route, uterine tachysystole, side effects, subject comfort, and other fetal heart rate disturbances. Cost, patient satisfaction, and neonatal outcomes were not addressed.

RESULTS: The groups were similar at baseline. There was no difference between the groups in change of Bishop score, preinduction cervical ripening times, total induction times, or mode of delivery. However, misoprostol-treated women were statistically more likely to have uterine contractile abnormalities (P <.001; number needed to harm [NNH]=5) and meconium passage (P=.01; NNH=6). Two women in the misoprostol group had hyperstimulation and required cesarean deliveries, but the infants did well. The most frequent complaint in the Foley group was mild discomfort at insertion, but there was no difference between the groups in overall maternal discomfort.

BACKGROUND: Pharmacologic methods of cervical ripening are associated with uterine contractile abnormalities. This randomized trial compares intravaginal misoprostol to transcervical Foley catheters for preinduction cervical ripening.

POPULATION STUDIED: A total of 111 pregnant women were admitted to a tertiary hospital for induction. Subjects had singleton vertex pregnancies at more than 28 weeks’ gestation and a Bishop score of less than 6; indications for induction included preeclampsia, oligohydramnios, postdates, and growth restriction. Exclusion criteria included previous uterine surgeries, rupture of membranes, previous induction, or use of a pre-induction agent during the pregnancy. Mean age was 26 years; 70% were nulliparous. Labor was managed actively with rupture of membranes as soon as possible, use of oxytocin, and an epidural rate of almost 90%. Although the study population included premature infants and the setting was different from that in which most family physicians deliver babies, the results probably apply to patients cared for by family physicians.

STUDY DESIGN AND VALIDITY: The participants were randomized with concealed allocation to placement of a transcervical Foley catheter or 50 mg intravaginal misoprostol. A 16F Foley catheter was inserted under direct visualization during a sterile speculum examination without an obturator, tenaculum, or cervical cleansing. Once past the internal os, the balloon was filled with 30 cc of sterile water, and gentle traction was applied (the end of the catheter was taped to the patient’s medial knee or thigh). Every 6 hours the catheter was adjusted to continue traction. In the misoprostol group, 50 μg was placed in the posterior fornix of the vagina every 4 hours up to 6 times. Oxytocin was begun after extrusion of the catheter or 4 hours after the last dose of misoprostol if no labor ensued or if all 6 doses had been given. Bishop scores were assigned before randomization and at the time of Foley catheter extrusion or last dose of misoprostol. The Student t test, Fisher exact test, Mann-Whitney U, and chi square were used to compare groups; no information was given about whether testing was by intention to treat. The methodologic strength of this study was adequate. Its major strengths were randomization with concealed allocation and adequate power. Major weaknesses included the lack of attention to potential confounding factors, such as labor support, that are known to influence the course of labor and the inability to mask the physician assessing the Bishop scores. Other minor weaknesses included the intrinsic lack of intra- and inter-observer reliability of the Bishop score, the possible variation in the timing of giving the Bishop score for the misoprostol group, and the lack of power for assessment of outcomes other than the primary ones.

OUTCOMES MEASURED: The primary outcome measure was change in Bishop score. Secondary outcome measures included length of time for preinduction cervical ripening, total time for induction, delivery route, uterine tachysystole, side effects, subject comfort, and other fetal heart rate disturbances. Cost, patient satisfaction, and neonatal outcomes were not addressed.

RESULTS: The groups were similar at baseline. There was no difference between the groups in change of Bishop score, preinduction cervical ripening times, total induction times, or mode of delivery. However, misoprostol-treated women were statistically more likely to have uterine contractile abnormalities (P <.001; number needed to harm [NNH]=5) and meconium passage (P=.01; NNH=6). Two women in the misoprostol group had hyperstimulation and required cesarean deliveries, but the infants did well. The most frequent complaint in the Foley group was mild discomfort at insertion, but there was no difference between the groups in overall maternal discomfort.

BACKGROUND: Pharmacologic methods of cervical ripening are associated with uterine contractile abnormalities. This randomized trial compares intravaginal misoprostol to transcervical Foley catheters for preinduction cervical ripening.

POPULATION STUDIED: A total of 111 pregnant women were admitted to a tertiary hospital for induction. Subjects had singleton vertex pregnancies at more than 28 weeks’ gestation and a Bishop score of less than 6; indications for induction included preeclampsia, oligohydramnios, postdates, and growth restriction. Exclusion criteria included previous uterine surgeries, rupture of membranes, previous induction, or use of a pre-induction agent during the pregnancy. Mean age was 26 years; 70% were nulliparous. Labor was managed actively with rupture of membranes as soon as possible, use of oxytocin, and an epidural rate of almost 90%. Although the study population included premature infants and the setting was different from that in which most family physicians deliver babies, the results probably apply to patients cared for by family physicians.

STUDY DESIGN AND VALIDITY: The participants were randomized with concealed allocation to placement of a transcervical Foley catheter or 50 mg intravaginal misoprostol. A 16F Foley catheter was inserted under direct visualization during a sterile speculum examination without an obturator, tenaculum, or cervical cleansing. Once past the internal os, the balloon was filled with 30 cc of sterile water, and gentle traction was applied (the end of the catheter was taped to the patient’s medial knee or thigh). Every 6 hours the catheter was adjusted to continue traction. In the misoprostol group, 50 μg was placed in the posterior fornix of the vagina every 4 hours up to 6 times. Oxytocin was begun after extrusion of the catheter or 4 hours after the last dose of misoprostol if no labor ensued or if all 6 doses had been given. Bishop scores were assigned before randomization and at the time of Foley catheter extrusion or last dose of misoprostol. The Student t test, Fisher exact test, Mann-Whitney U, and chi square were used to compare groups; no information was given about whether testing was by intention to treat. The methodologic strength of this study was adequate. Its major strengths were randomization with concealed allocation and adequate power. Major weaknesses included the lack of attention to potential confounding factors, such as labor support, that are known to influence the course of labor and the inability to mask the physician assessing the Bishop scores. Other minor weaknesses included the intrinsic lack of intra- and inter-observer reliability of the Bishop score, the possible variation in the timing of giving the Bishop score for the misoprostol group, and the lack of power for assessment of outcomes other than the primary ones.

OUTCOMES MEASURED: The primary outcome measure was change in Bishop score. Secondary outcome measures included length of time for preinduction cervical ripening, total time for induction, delivery route, uterine tachysystole, side effects, subject comfort, and other fetal heart rate disturbances. Cost, patient satisfaction, and neonatal outcomes were not addressed.

RESULTS: The groups were similar at baseline. There was no difference between the groups in change of Bishop score, preinduction cervical ripening times, total induction times, or mode of delivery. However, misoprostol-treated women were statistically more likely to have uterine contractile abnormalities (P <.001; number needed to harm [NNH]=5) and meconium passage (P=.01; NNH=6). Two women in the misoprostol group had hyperstimulation and required cesarean deliveries, but the infants did well. The most frequent complaint in the Foley group was mild discomfort at insertion, but there was no difference between the groups in overall maternal discomfort.