Miriam E. Tucker

Major Finding: During the first stage of labor, mean verbal analog scale pain score was significantly less in the SE group compared with EA (1.36 vs. 1.89), but the difference was not significant by the end of the second stage.

Data Source: A randomized, controlled trial of 800 women.

Disclosures: Dr. Gambling said he had no financial disclosures to report.

SAN ANTONIO — Combined spinal-epidural anesthesia was superior to traditional epidural for first-stage anesthesia but there were no differences in second stage or in delivery pain in a randomized, controlled comparison of the two methods among 800 women.

The Epidural Analgesia and Spinal Epidural Analgesia (EASE) study also showed that concerns about epidurals failing with combined spinal-epidural (SE) because of the inability to provide a test dose are unfounded, Dr. David R. Gambling reported.

Previous studies comparing the techniques have had mixed results. A Cochrane review showed that CSE had less rescue analgesia and less urinary retention but more pruritis (Cochrane Database Syst. Rev. 2007 [doi:10.1002/14651858.CD003401.pub2]). Compared with low-dose epidural anesthesia (EA), combined SE had faster-onset analgesia, more pruritis, and lower umbilical cord artery pH, but there was no mention of progress of cervical dilation, noted Dr. Gambling of the Sharp Mary Birch Hospital for Women and Newborns and the University of California, San Diego.

In EASE, 398 women received EA, consisting of 10 mL 0.125% bupivacaine with 2 mcg/mL fentanyl in two 5-mL doses via epidural needle, followed by 5 mL of the same solution via epidural catheter (total dose 15 mL). The 402 in the SE group were given 2.5 mL 0.125% isobaric bupivacaine plus 2 mcg/mL fentanyl via 26-g GM spinal needle prior to epidural catheter placement.

In both groups, medications were administered at first request for neuraxial anesthesia. Labor was managed by registered nurses and obstetricians who were blinded to group assignment.

There were no significant differences between the groups in age, height, weight, body mass index, estimated gestational age, cervical dilation at epidural insertion, or pre-epidural verbal analog scale (VAS) pain scores. However, the time to complete analgesia (from initial EA and SE injection until patient reported VAS scores of 0 or 1 was significantly less with the SE group, 11 vs. 22 minutes.

The second stage of labor was statistically significantly shorter with EA (68 vs. 78 minutes), but the difference may not be clinically significant. There were no significant differences in time from epidural induction until cervical dilation reached 10 cm, duration of pushing, or rate of cervical dilation. There were also no differences in the use of instrumentation with vaginal delivery or need for cesarean section.

During the first stage of labor, the mean VAS pain score was significantly less in the SE group. compared with EA (1.36 vs. 1.89) and also at 1 hour of labor (0.26 vs. 0.72), despite a slightly lower rate of patient-controlled analgesia use during the first stage (10 vs. 11 mL/hr). The proportion of women with mean VAS scores of zero at the end of stage 1 was significantly higher with SE (42% vs. 31% with EA), but the difference was not significant by the end of the second stage, he said.

Need for epidural top-up was greater in the EA group (26% vs. 16%), as was the need for more than one top-up (21% vs. 9%). Only a small proportion of each group (2% EA and 1.2% SE) required replacement of the epidural catheter, suggesting that there should not be concern about epidurals failing with SE because of inability to provide a test dose, he commented.

Fetal heart rate decelerations within 30 minutes of analgesic induction were more common in the SE group (8.5% vs. 4.5%), but none required emergency c-section. The proportions with Apgar scores below 7 at 1 and 5 minutes were less than 5% and less than 0.5%, respectively, in both groups.

Patient satisfaction with their mode of analgesia did not differ, at 98% for SE and 96% for EA, Dr. Gambling reported.

Patient satisfaction with their mode of analgesia did not differ between the two groups.

Source DR. GAMBLING

Major Finding: During the first stage of labor, mean verbal analog scale pain score was significantly less in the SE group compared with EA (1.36 vs. 1.89), but the difference was not significant by the end of the second stage.

Data Source: A randomized, controlled trial of 800 women.

Disclosures: Dr. Gambling said he had no financial disclosures to report.

SAN ANTONIO — Combined spinal-epidural anesthesia was superior to traditional epidural for first-stage anesthesia but there were no differences in second stage or in delivery pain in a randomized, controlled comparison of the two methods among 800 women.

The Epidural Analgesia and Spinal Epidural Analgesia (EASE) study also showed that concerns about epidurals failing with combined spinal-epidural (SE) because of the inability to provide a test dose are unfounded, Dr. David R. Gambling reported.

Previous studies comparing the techniques have had mixed results. A Cochrane review showed that CSE had less rescue analgesia and less urinary retention but more pruritis (Cochrane Database Syst. Rev. 2007 [doi:10.1002/14651858.CD003401.pub2]). Compared with low-dose epidural anesthesia (EA), combined SE had faster-onset analgesia, more pruritis, and lower umbilical cord artery pH, but there was no mention of progress of cervical dilation, noted Dr. Gambling of the Sharp Mary Birch Hospital for Women and Newborns and the University of California, San Diego.

In EASE, 398 women received EA, consisting of 10 mL 0.125% bupivacaine with 2 mcg/mL fentanyl in two 5-mL doses via epidural needle, followed by 5 mL of the same solution via epidural catheter (total dose 15 mL). The 402 in the SE group were given 2.5 mL 0.125% isobaric bupivacaine plus 2 mcg/mL fentanyl via 26-g GM spinal needle prior to epidural catheter placement.

In both groups, medications were administered at first request for neuraxial anesthesia. Labor was managed by registered nurses and obstetricians who were blinded to group assignment.

There were no significant differences between the groups in age, height, weight, body mass index, estimated gestational age, cervical dilation at epidural insertion, or pre-epidural verbal analog scale (VAS) pain scores. However, the time to complete analgesia (from initial EA and SE injection until patient reported VAS scores of 0 or 1 was significantly less with the SE group, 11 vs. 22 minutes.

The second stage of labor was statistically significantly shorter with EA (68 vs. 78 minutes), but the difference may not be clinically significant. There were no significant differences in time from epidural induction until cervical dilation reached 10 cm, duration of pushing, or rate of cervical dilation. There were also no differences in the use of instrumentation with vaginal delivery or need for cesarean section.

During the first stage of labor, the mean VAS pain score was significantly less in the SE group. compared with EA (1.36 vs. 1.89) and also at 1 hour of labor (0.26 vs. 0.72), despite a slightly lower rate of patient-controlled analgesia use during the first stage (10 vs. 11 mL/hr). The proportion of women with mean VAS scores of zero at the end of stage 1 was significantly higher with SE (42% vs. 31% with EA), but the difference was not significant by the end of the second stage, he said.

Need for epidural top-up was greater in the EA group (26% vs. 16%), as was the need for more than one top-up (21% vs. 9%). Only a small proportion of each group (2% EA and 1.2% SE) required replacement of the epidural catheter, suggesting that there should not be concern about epidurals failing with SE because of inability to provide a test dose, he commented.

Fetal heart rate decelerations within 30 minutes of analgesic induction were more common in the SE group (8.5% vs. 4.5%), but none required emergency c-section. The proportions with Apgar scores below 7 at 1 and 5 minutes were less than 5% and less than 0.5%, respectively, in both groups.

Patient satisfaction with their mode of analgesia did not differ, at 98% for SE and 96% for EA, Dr. Gambling reported.

Patient satisfaction with their mode of analgesia did not differ between the two groups.

Source DR. GAMBLING

Major Finding: During the first stage of labor, mean verbal analog scale pain score was significantly less in the SE group compared with EA (1.36 vs. 1.89), but the difference was not significant by the end of the second stage.

Data Source: A randomized, controlled trial of 800 women.

Disclosures: Dr. Gambling said he had no financial disclosures to report.

SAN ANTONIO — Combined spinal-epidural anesthesia was superior to traditional epidural for first-stage anesthesia but there were no differences in second stage or in delivery pain in a randomized, controlled comparison of the two methods among 800 women.

The Epidural Analgesia and Spinal Epidural Analgesia (EASE) study also showed that concerns about epidurals failing with combined spinal-epidural (SE) because of the inability to provide a test dose are unfounded, Dr. David R. Gambling reported.

Previous studies comparing the techniques have had mixed results. A Cochrane review showed that CSE had less rescue analgesia and less urinary retention but more pruritis (Cochrane Database Syst. Rev. 2007 [doi:10.1002/14651858.CD003401.pub2]). Compared with low-dose epidural anesthesia (EA), combined SE had faster-onset analgesia, more pruritis, and lower umbilical cord artery pH, but there was no mention of progress of cervical dilation, noted Dr. Gambling of the Sharp Mary Birch Hospital for Women and Newborns and the University of California, San Diego.

In EASE, 398 women received EA, consisting of 10 mL 0.125% bupivacaine with 2 mcg/mL fentanyl in two 5-mL doses via epidural needle, followed by 5 mL of the same solution via epidural catheter (total dose 15 mL). The 402 in the SE group were given 2.5 mL 0.125% isobaric bupivacaine plus 2 mcg/mL fentanyl via 26-g GM spinal needle prior to epidural catheter placement.

In both groups, medications were administered at first request for neuraxial anesthesia. Labor was managed by registered nurses and obstetricians who were blinded to group assignment.

There were no significant differences between the groups in age, height, weight, body mass index, estimated gestational age, cervical dilation at epidural insertion, or pre-epidural verbal analog scale (VAS) pain scores. However, the time to complete analgesia (from initial EA and SE injection until patient reported VAS scores of 0 or 1 was significantly less with the SE group, 11 vs. 22 minutes.

The second stage of labor was statistically significantly shorter with EA (68 vs. 78 minutes), but the difference may not be clinically significant. There were no significant differences in time from epidural induction until cervical dilation reached 10 cm, duration of pushing, or rate of cervical dilation. There were also no differences in the use of instrumentation with vaginal delivery or need for cesarean section.

During the first stage of labor, the mean VAS pain score was significantly less in the SE group. compared with EA (1.36 vs. 1.89) and also at 1 hour of labor (0.26 vs. 0.72), despite a slightly lower rate of patient-controlled analgesia use during the first stage (10 vs. 11 mL/hr). The proportion of women with mean VAS scores of zero at the end of stage 1 was significantly higher with SE (42% vs. 31% with EA), but the difference was not significant by the end of the second stage, he said.

Need for epidural top-up was greater in the EA group (26% vs. 16%), as was the need for more than one top-up (21% vs. 9%). Only a small proportion of each group (2% EA and 1.2% SE) required replacement of the epidural catheter, suggesting that there should not be concern about epidurals failing with SE because of inability to provide a test dose, he commented.

Fetal heart rate decelerations within 30 minutes of analgesic induction were more common in the SE group (8.5% vs. 4.5%), but none required emergency c-section. The proportions with Apgar scores below 7 at 1 and 5 minutes were less than 5% and less than 0.5%, respectively, in both groups.

Patient satisfaction with their mode of analgesia did not differ, at 98% for SE and 96% for EA, Dr. Gambling reported.

Patient satisfaction with their mode of analgesia did not differ between the two groups.

Source DR. GAMBLING

Major Finding: Total bupivacaine consumption per hour of analgesia was significantly lower among patients who received 10 mL over 60 minutes (10.3 mg/hour), compared with 11.3 mg/hour for those receiving 2.5 mL/15 minutes and 11.1 mg/hour for 5 mL/30 minutes.

Data Source: Randomized, controlled, double-blinded trial of 180 laboring women.

Disclosures: Dr. Wong said she had no financial conflicts of interest.

SAN ANTONIO — Providing epidural anesthesia in programmed boluses of higher volume with a longer duration between doses decreased total anesthetic consumption and variability without decreasing patient satisfaction in a randomized, controlled, double-blinded study of 180 laboring women.

Increasing evidence suggests that delivery of epidural anesthesia via intermittent bolus provides more effective anesthesia than does continuous infusion, said Dr. Cynthia A. Wong of Northwestern University, Chicago.

In a previous study, Dr. Wong and her colleagues reported that the currently available pumps used for patient-controlled epidural anesthesia (PCEA) also can be programmed to automatically deliver boluses at regular intervals, and that this “programmed intermittent epidural bolus” (PIEB) resulted in similar analgesia but with a smaller bupivacaine dose and better patient satisfaction, compared with continuous epidural infusion (CEI) for maintenance of epidural labor analgesia (Anesth. Analg. 2006;102:904-9).

As a follow-up, the current study investigated the effect of specific combinations of bolus volumes and time intervals to determine which is optimal. The subjects were healthy nulliparas with cervical dilation 2-5 cm. All received combined spinal-epidural anesthesia comprising intrathecal bupivacaine 1.25 mg/fentanyl 15 mcg and a test dose of epidural lidocaine 45 mg/epinephrine 15 mcg. The epidural maintenance solution consisted of bupivacaine 0.0625% with fentanyl 2 mcg/mL. Breakthrough pain was treated with PCEA and if needed, a manual bolus dose by the anesthesiologist.

The maintenance epidural technique was initiated 15 minutes after the intrathecal injection. Patients were randomized to one of three groups: 66 received 2.5 mL by the pump every 15 minutes (2.5/15), 60 received 5 mL every 30 minutes (5/30), and 54 got 10 mL every 60 minutes (10/60). Thus, all patients received the same total volume of drug but it was distributed differently, Dr. Wong noted.

All of the women had successful analgesia, and there were no differences in maximum oxytocin dose or mode of delivery among the groups.

The primary outcome, total bupivacaine consumption per hour of analgesia, was significantly lower in the 10/60 group compared with the other two, with a mean of 10.3 mg/hr versus 11.3 mg/hr for the 2.5/15 patients and 11.1 mg/hr with 5/30. There was also less variability in dosing from hour to hour in the 10/60 group, she noted.

There were no significant differences in any secondary variable, including Visual Analog Pain score, motor block (Bromage greater than 0), number of PCEA requests, time to first request for manual bolus, number of subjects requiring manual bolus, patient satisfaction score, or extent of sensory blockade, as measured by both cold stimulus and von Frey hair threshold tests.

The mechanism isn't entirely clear. All studies of PIEB have shown that the technique provides equal or better analgesia than does CIE with a lower dose of drug. But, if as hypothesized, the reason is that boluses provides better spread in the epidural space, then it is “interesting” that this study found no difference in the extent of sensory blockade among the three groups. Indeed, data on the extent of sensory blockade in other studies of PIEB have been inconsistent, she said.

Other variables, such as differences in catheter design or patient demographics, might also contribute to the variability in extent of analgesia, she added.

In response to a question from the audience about whether these findings have changed her clinical practice, Dr. Wong noted that there is currently no commercially available pump that delivers both PCEA and PIEB.

However, her institution has “considerably backed off using continuous infusion rate” and now relies more on patient-controlled bolusing, resulting in a lower manual re-bolus rate. “There's very solid evidence that giving the drug as a bolus, by whatever means—by the patient, the machine, or the anesthesiologist—is a more efficient technique.

There is currently no commercially available pump that delivers both PCEA and PIEB.

Source DR. WONG

Major Finding: Total bupivacaine consumption per hour of analgesia was significantly lower among patients who received 10 mL over 60 minutes (10.3 mg/hour), compared with 11.3 mg/hour for those receiving 2.5 mL/15 minutes and 11.1 mg/hour for 5 mL/30 minutes.

Data Source: Randomized, controlled, double-blinded trial of 180 laboring women.

Disclosures: Dr. Wong said she had no financial conflicts of interest.

SAN ANTONIO — Providing epidural anesthesia in programmed boluses of higher volume with a longer duration between doses decreased total anesthetic consumption and variability without decreasing patient satisfaction in a randomized, controlled, double-blinded study of 180 laboring women.

Increasing evidence suggests that delivery of epidural anesthesia via intermittent bolus provides more effective anesthesia than does continuous infusion, said Dr. Cynthia A. Wong of Northwestern University, Chicago.

In a previous study, Dr. Wong and her colleagues reported that the currently available pumps used for patient-controlled epidural anesthesia (PCEA) also can be programmed to automatically deliver boluses at regular intervals, and that this “programmed intermittent epidural bolus” (PIEB) resulted in similar analgesia but with a smaller bupivacaine dose and better patient satisfaction, compared with continuous epidural infusion (CEI) for maintenance of epidural labor analgesia (Anesth. Analg. 2006;102:904-9).

As a follow-up, the current study investigated the effect of specific combinations of bolus volumes and time intervals to determine which is optimal. The subjects were healthy nulliparas with cervical dilation 2-5 cm. All received combined spinal-epidural anesthesia comprising intrathecal bupivacaine 1.25 mg/fentanyl 15 mcg and a test dose of epidural lidocaine 45 mg/epinephrine 15 mcg. The epidural maintenance solution consisted of bupivacaine 0.0625% with fentanyl 2 mcg/mL. Breakthrough pain was treated with PCEA and if needed, a manual bolus dose by the anesthesiologist.

The maintenance epidural technique was initiated 15 minutes after the intrathecal injection. Patients were randomized to one of three groups: 66 received 2.5 mL by the pump every 15 minutes (2.5/15), 60 received 5 mL every 30 minutes (5/30), and 54 got 10 mL every 60 minutes (10/60). Thus, all patients received the same total volume of drug but it was distributed differently, Dr. Wong noted.

All of the women had successful analgesia, and there were no differences in maximum oxytocin dose or mode of delivery among the groups.

The primary outcome, total bupivacaine consumption per hour of analgesia, was significantly lower in the 10/60 group compared with the other two, with a mean of 10.3 mg/hr versus 11.3 mg/hr for the 2.5/15 patients and 11.1 mg/hr with 5/30. There was also less variability in dosing from hour to hour in the 10/60 group, she noted.

There were no significant differences in any secondary variable, including Visual Analog Pain score, motor block (Bromage greater than 0), number of PCEA requests, time to first request for manual bolus, number of subjects requiring manual bolus, patient satisfaction score, or extent of sensory blockade, as measured by both cold stimulus and von Frey hair threshold tests.

The mechanism isn't entirely clear. All studies of PIEB have shown that the technique provides equal or better analgesia than does CIE with a lower dose of drug. But, if as hypothesized, the reason is that boluses provides better spread in the epidural space, then it is “interesting” that this study found no difference in the extent of sensory blockade among the three groups. Indeed, data on the extent of sensory blockade in other studies of PIEB have been inconsistent, she said.

Other variables, such as differences in catheter design or patient demographics, might also contribute to the variability in extent of analgesia, she added.

In response to a question from the audience about whether these findings have changed her clinical practice, Dr. Wong noted that there is currently no commercially available pump that delivers both PCEA and PIEB.

However, her institution has “considerably backed off using continuous infusion rate” and now relies more on patient-controlled bolusing, resulting in a lower manual re-bolus rate. “There's very solid evidence that giving the drug as a bolus, by whatever means—by the patient, the machine, or the anesthesiologist—is a more efficient technique.

There is currently no commercially available pump that delivers both PCEA and PIEB.

Source DR. WONG

Major Finding: Total bupivacaine consumption per hour of analgesia was significantly lower among patients who received 10 mL over 60 minutes (10.3 mg/hour), compared with 11.3 mg/hour for those receiving 2.5 mL/15 minutes and 11.1 mg/hour for 5 mL/30 minutes.

Data Source: Randomized, controlled, double-blinded trial of 180 laboring women.

Disclosures: Dr. Wong said she had no financial conflicts of interest.

SAN ANTONIO — Providing epidural anesthesia in programmed boluses of higher volume with a longer duration between doses decreased total anesthetic consumption and variability without decreasing patient satisfaction in a randomized, controlled, double-blinded study of 180 laboring women.

Increasing evidence suggests that delivery of epidural anesthesia via intermittent bolus provides more effective anesthesia than does continuous infusion, said Dr. Cynthia A. Wong of Northwestern University, Chicago.

In a previous study, Dr. Wong and her colleagues reported that the currently available pumps used for patient-controlled epidural anesthesia (PCEA) also can be programmed to automatically deliver boluses at regular intervals, and that this “programmed intermittent epidural bolus” (PIEB) resulted in similar analgesia but with a smaller bupivacaine dose and better patient satisfaction, compared with continuous epidural infusion (CEI) for maintenance of epidural labor analgesia (Anesth. Analg. 2006;102:904-9).

As a follow-up, the current study investigated the effect of specific combinations of bolus volumes and time intervals to determine which is optimal. The subjects were healthy nulliparas with cervical dilation 2-5 cm. All received combined spinal-epidural anesthesia comprising intrathecal bupivacaine 1.25 mg/fentanyl 15 mcg and a test dose of epidural lidocaine 45 mg/epinephrine 15 mcg. The epidural maintenance solution consisted of bupivacaine 0.0625% with fentanyl 2 mcg/mL. Breakthrough pain was treated with PCEA and if needed, a manual bolus dose by the anesthesiologist.

The maintenance epidural technique was initiated 15 minutes after the intrathecal injection. Patients were randomized to one of three groups: 66 received 2.5 mL by the pump every 15 minutes (2.5/15), 60 received 5 mL every 30 minutes (5/30), and 54 got 10 mL every 60 minutes (10/60). Thus, all patients received the same total volume of drug but it was distributed differently, Dr. Wong noted.

All of the women had successful analgesia, and there were no differences in maximum oxytocin dose or mode of delivery among the groups.

The primary outcome, total bupivacaine consumption per hour of analgesia, was significantly lower in the 10/60 group compared with the other two, with a mean of 10.3 mg/hr versus 11.3 mg/hr for the 2.5/15 patients and 11.1 mg/hr with 5/30. There was also less variability in dosing from hour to hour in the 10/60 group, she noted.

There were no significant differences in any secondary variable, including Visual Analog Pain score, motor block (Bromage greater than 0), number of PCEA requests, time to first request for manual bolus, number of subjects requiring manual bolus, patient satisfaction score, or extent of sensory blockade, as measured by both cold stimulus and von Frey hair threshold tests.

The mechanism isn't entirely clear. All studies of PIEB have shown that the technique provides equal or better analgesia than does CIE with a lower dose of drug. But, if as hypothesized, the reason is that boluses provides better spread in the epidural space, then it is “interesting” that this study found no difference in the extent of sensory blockade among the three groups. Indeed, data on the extent of sensory blockade in other studies of PIEB have been inconsistent, she said.

Other variables, such as differences in catheter design or patient demographics, might also contribute to the variability in extent of analgesia, she added.

In response to a question from the audience about whether these findings have changed her clinical practice, Dr. Wong noted that there is currently no commercially available pump that delivers both PCEA and PIEB.

However, her institution has “considerably backed off using continuous infusion rate” and now relies more on patient-controlled bolusing, resulting in a lower manual re-bolus rate. “There's very solid evidence that giving the drug as a bolus, by whatever means—by the patient, the machine, or the anesthesiologist—is a more efficient technique.

There is currently no commercially available pump that delivers both PCEA and PIEB.

Source DR. WONG

Major Finding: At 6 months, mean reduction in hemoglobin A_1c was similar between patients taking VIAject and those taking regular human insulin before meals but hypoglycemic event rates were significantly reduced with VIAject (0.33 vs. 0.66 events/month) and weight gain was significantly less with VIAject (0.46 vs. 1.35 kg).

Data Source: Open-label, multicenter, randomized phase III trial of 471 patients with type 2 diabetes

Disclosures: Biodel Corp. funded the study. Dr. Rodbard serves on the company's advisory board and has received research support from it. She also has consultant, speaker, and/or research grant support relationships with other companies that make diabetes-related products.

Orlando — An investigational rapid-acting human insulin formulation provided similar glucose control to regular human insulin but with a twofold reduction in hypoglycemia and significantly less weight gain in a 6-month, multicenter, open-label study of 471 patients with type 2 diabetes.

The insulin formulation, called VIAject, is absorbed more rapidly after subcutaneous injection than either insulin lispro or regular human insulin. Findings from Biodel Corp.'s phase III study of VIAject were presented by Dr. Helena Rodbard, an endocrinologist in private practice in Rockville, Md., and a past president of the American College of Endocrinology and the American Association of Clinical Endocrinologists.

The subjects had a mean age of 56 years, slightly more than half were male, and the mean body mass index was 33 kg/m

At 6 months, the mean reduction in hemoglobin A_1c was similar in the two groups, with the VIAject group dropping by 0.56 percentage points and the regular human insulin group by 0.70 points. But nonsevere hypoglycemic event rates were significantly reduced in the patients taking VIAject, at just 0.33 events/month, compared with 0.66 events/month with regular human insulin, Dr. Rodbard reported.

Patients taking VIAject also gained significantly less weight, an average of 0.46 kg vs. 1.35 kg with regular human insulin.

Insulin antibody levels and other laboratory tests monitoring safety were similar for both groups. Injection site pain or irritation was greater with VIAject, but this declined during the course of the study. Moreover, the proposed U-100 pH-neutral commercial formulation of VIAject is associated with less injection site discomfort than the U-25 pH 4 version used in this study, she pointed out.

Biodel is seeking Food and Drug Administration clearance to market VIAject based on two pivotal 6-month phase III clinical trials in patients with type 1 and type 2 diabetes, as well as results from long-term, 18-month safety extension trials for patients who completed the two pivotal phase III clinical trials. The Prescription Drug User Fee Act action date for Biodel's new drug application is expected to be Oct. 30, 2010, the company said in a statement.

Major Finding: At 6 months, mean reduction in hemoglobin A_1c was similar between patients taking VIAject and those taking regular human insulin before meals but hypoglycemic event rates were significantly reduced with VIAject (0.33 vs. 0.66 events/month) and weight gain was significantly less with VIAject (0.46 vs. 1.35 kg).

Data Source: Open-label, multicenter, randomized phase III trial of 471 patients with type 2 diabetes

Disclosures: Biodel Corp. funded the study. Dr. Rodbard serves on the company's advisory board and has received research support from it. She also has consultant, speaker, and/or research grant support relationships with other companies that make diabetes-related products.

Orlando — An investigational rapid-acting human insulin formulation provided similar glucose control to regular human insulin but with a twofold reduction in hypoglycemia and significantly less weight gain in a 6-month, multicenter, open-label study of 471 patients with type 2 diabetes.

The insulin formulation, called VIAject, is absorbed more rapidly after subcutaneous injection than either insulin lispro or regular human insulin. Findings from Biodel Corp.'s phase III study of VIAject were presented by Dr. Helena Rodbard, an endocrinologist in private practice in Rockville, Md., and a past president of the American College of Endocrinology and the American Association of Clinical Endocrinologists.

The subjects had a mean age of 56 years, slightly more than half were male, and the mean body mass index was 33 kg/m

At 6 months, the mean reduction in hemoglobin A_1c was similar in the two groups, with the VIAject group dropping by 0.56 percentage points and the regular human insulin group by 0.70 points. But nonsevere hypoglycemic event rates were significantly reduced in the patients taking VIAject, at just 0.33 events/month, compared with 0.66 events/month with regular human insulin, Dr. Rodbard reported.

Patients taking VIAject also gained significantly less weight, an average of 0.46 kg vs. 1.35 kg with regular human insulin.

Insulin antibody levels and other laboratory tests monitoring safety were similar for both groups. Injection site pain or irritation was greater with VIAject, but this declined during the course of the study. Moreover, the proposed U-100 pH-neutral commercial formulation of VIAject is associated with less injection site discomfort than the U-25 pH 4 version used in this study, she pointed out.

Biodel is seeking Food and Drug Administration clearance to market VIAject based on two pivotal 6-month phase III clinical trials in patients with type 1 and type 2 diabetes, as well as results from long-term, 18-month safety extension trials for patients who completed the two pivotal phase III clinical trials. The Prescription Drug User Fee Act action date for Biodel's new drug application is expected to be Oct. 30, 2010, the company said in a statement.

Major Finding: At 6 months, mean reduction in hemoglobin A_1c was similar between patients taking VIAject and those taking regular human insulin before meals but hypoglycemic event rates were significantly reduced with VIAject (0.33 vs. 0.66 events/month) and weight gain was significantly less with VIAject (0.46 vs. 1.35 kg).

Data Source: Open-label, multicenter, randomized phase III trial of 471 patients with type 2 diabetes

Disclosures: Biodel Corp. funded the study. Dr. Rodbard serves on the company's advisory board and has received research support from it. She also has consultant, speaker, and/or research grant support relationships with other companies that make diabetes-related products.

Orlando — An investigational rapid-acting human insulin formulation provided similar glucose control to regular human insulin but with a twofold reduction in hypoglycemia and significantly less weight gain in a 6-month, multicenter, open-label study of 471 patients with type 2 diabetes.

The insulin formulation, called VIAject, is absorbed more rapidly after subcutaneous injection than either insulin lispro or regular human insulin. Findings from Biodel Corp.'s phase III study of VIAject were presented by Dr. Helena Rodbard, an endocrinologist in private practice in Rockville, Md., and a past president of the American College of Endocrinology and the American Association of Clinical Endocrinologists.

The subjects had a mean age of 56 years, slightly more than half were male, and the mean body mass index was 33 kg/m

At 6 months, the mean reduction in hemoglobin A_1c was similar in the two groups, with the VIAject group dropping by 0.56 percentage points and the regular human insulin group by 0.70 points. But nonsevere hypoglycemic event rates were significantly reduced in the patients taking VIAject, at just 0.33 events/month, compared with 0.66 events/month with regular human insulin, Dr. Rodbard reported.

Patients taking VIAject also gained significantly less weight, an average of 0.46 kg vs. 1.35 kg with regular human insulin.

Insulin antibody levels and other laboratory tests monitoring safety were similar for both groups. Injection site pain or irritation was greater with VIAject, but this declined during the course of the study. Moreover, the proposed U-100 pH-neutral commercial formulation of VIAject is associated with less injection site discomfort than the U-25 pH 4 version used in this study, she pointed out.

Biodel is seeking Food and Drug Administration clearance to market VIAject based on two pivotal 6-month phase III clinical trials in patients with type 1 and type 2 diabetes, as well as results from long-term, 18-month safety extension trials for patients who completed the two pivotal phase III clinical trials. The Prescription Drug User Fee Act action date for Biodel's new drug application is expected to be Oct. 30, 2010, the company said in a statement.

Major Finding: One or more hypoglycemic events occurred in 26 of 40 patients despite hemoglobin A_1c values above 8%. Of 102 hypoglycemic episodes, 95 were not recognized by fingerstick testing or by symptoms.

Data Source: Study using blinded continuous glucose monitoring for 72 or more hours in 40 community-dwelling diabetes patients seen at a tertiary diabetes center.

Disclosures: The American Diabetes Association funded the study. Dr. Munshi stated that she had no conflicts of interest.

Orlando — Hypoglycemic episodes were common despite high hemoglobin A_1c levels among 40 elderly community-living adults with diabetes who underwent continuous glucose monitoring.

“Raising A_1c goals may not be adequate to prevent hypoglycemia in this vulnerable population,” Dr. Medha N. Munshi said.

Guidelines from the American Diabetes Association and the American Geriatric Society advise that the usual recommended hemoglobin A_1c target of less than 6.5%-7.0% might be relaxed for elderly adults who have a history of severe hypoglycemia, limited life expectancy, advanced diabetes complications, or extensive comorbidity (Diabetes Care 2010;33 [suppl 1]).

In practice, this has been interpreted as a goal of less than 8%. The ADA guideline was based on level “C” evidence, and no study has ever clearly demonstrated that raising the HbA_1c target actually reduces the risk of hypoglycemia, said Dr. Munshi, director of the Joslin Diabetes Center geriatric programs at Beth Israel Deaconess Medical Center, Boston.

The current study used blinded continuous glucose monitoring (CGM) for 72 hours or longer in 40 community-dwelling diabetes patients who were seen at the Joslin Diabetes Center. To be included, they had to be older than 69 years and have a hemoglobin A_1c level greater than 8%.

The study group was 60% female and 80% white. Mean age was 75 years and mean HbA_1c was 9.3%. The patients took a mean of eight medications per day, with more than half (55%) on insulin alone, and another 38% on insulin plus one or more oral agents. Two-thirds had type 2 diabetes and the rest had type 1. Nearly a quarter were living alone, said Dr. Munshi, also with Harvard Medical School, Boston.

Patients performed four fingerstick glucose measurements per day and kept daily diaries of hypoglycemic symptoms, diet, and physical activity. One or more hypoglycemic events, defined as a glucose value less than 70 mg/dL, occurred in 26 of the 40 patients. Nearly three-fourths of the events involved glucose levels of 50-59 mg/dL, and in just under half the glucose dropped below 50 mg/dL.

The 26 patients experiencing hypoglycemia did not differ from the 14 without such events in patient characteristics including age, diabetes duration, HbA_1c, or insulin treatment. There were also no differences between those who did and did not have hypoglycemia in the comorbidities cognitive dysfunction, depression, falls in the past 6 months, number of medications, hypertension, or vision/hearing problems.

Of the 26 with hypoglycemia, 12 had HbA_1c levels above 9%. “Even a high A_1c doesn't preclude lows. Hemoglobin A_1c measures the mean. There are wide fluctuations in this population,” Dr. Munshi noted.

There were a total of 102 hypoglycemic episodes, with a mean duration of 3 hours per patient. Nocturnal episodes lasted for a mean of 2.5 hours. “The duration of episodes was quite concerning,” she said.

Surprisingly, more than half (58%) of the 102 hypoglycemic episodes occurred among the 16 patients with type 2 diabetes, with a mean duration of nocturnal hypoglycemia nearly twice that of the hypoglycemia in the type 1 patients (2.9 vs. 1.6 hours). “Even the type 2 patients had wide glycemic excursions,” Dr. Munshi commented.

Also of concern, the majority of episodes (95 of the 102) were not recognized by fingerstick testing or by the patients' symptoms. Moreover, there were no significant relationships between severity of hypoglycemia and age, type of diabetes, duration, HbA_1c, treatment, or living alone.

In a follow-up interview, Dr. Munshi cited a previous study she and her colleagues published last year, showing that simplification of complex insulin regimens by using C peptide to assess whether patients could eliminate or reduce the amount of insulin taken, and adding oral agents instead, reduced hypoglycemic episodes without deterioration of glycemic control (Am. J. Med. 2009;122:395-7).

“Elderly patients with other comorbidities are unable to follow complex insulin regimens appropriately and end up having wide fluctuations in their glucose values. If a treatment regimen is designed with consideration for an elderly patient's self-care abilities, risk of hypoglycemia can be reduced,” she said.

And when available, CGM can be extremely helpful for elderly patients, Dr. Munshi said.

Major Finding: One or more hypoglycemic events occurred in 26 of 40 patients despite hemoglobin A_1c values above 8%. Of 102 hypoglycemic episodes, 95 were not recognized by fingerstick testing or by symptoms.

Data Source: Study using blinded continuous glucose monitoring for 72 or more hours in 40 community-dwelling diabetes patients seen at a tertiary diabetes center.

Disclosures: The American Diabetes Association funded the study. Dr. Munshi stated that she had no conflicts of interest.

Orlando — Hypoglycemic episodes were common despite high hemoglobin A_1c levels among 40 elderly community-living adults with diabetes who underwent continuous glucose monitoring.

“Raising A_1c goals may not be adequate to prevent hypoglycemia in this vulnerable population,” Dr. Medha N. Munshi said.

Guidelines from the American Diabetes Association and the American Geriatric Society advise that the usual recommended hemoglobin A_1c target of less than 6.5%-7.0% might be relaxed for elderly adults who have a history of severe hypoglycemia, limited life expectancy, advanced diabetes complications, or extensive comorbidity (Diabetes Care 2010;33 [suppl 1]).

In practice, this has been interpreted as a goal of less than 8%. The ADA guideline was based on level “C” evidence, and no study has ever clearly demonstrated that raising the HbA_1c target actually reduces the risk of hypoglycemia, said Dr. Munshi, director of the Joslin Diabetes Center geriatric programs at Beth Israel Deaconess Medical Center, Boston.

The current study used blinded continuous glucose monitoring (CGM) for 72 hours or longer in 40 community-dwelling diabetes patients who were seen at the Joslin Diabetes Center. To be included, they had to be older than 69 years and have a hemoglobin A_1c level greater than 8%.

The study group was 60% female and 80% white. Mean age was 75 years and mean HbA_1c was 9.3%. The patients took a mean of eight medications per day, with more than half (55%) on insulin alone, and another 38% on insulin plus one or more oral agents. Two-thirds had type 2 diabetes and the rest had type 1. Nearly a quarter were living alone, said Dr. Munshi, also with Harvard Medical School, Boston.

Patients performed four fingerstick glucose measurements per day and kept daily diaries of hypoglycemic symptoms, diet, and physical activity. One or more hypoglycemic events, defined as a glucose value less than 70 mg/dL, occurred in 26 of the 40 patients. Nearly three-fourths of the events involved glucose levels of 50-59 mg/dL, and in just under half the glucose dropped below 50 mg/dL.

The 26 patients experiencing hypoglycemia did not differ from the 14 without such events in patient characteristics including age, diabetes duration, HbA_1c, or insulin treatment. There were also no differences between those who did and did not have hypoglycemia in the comorbidities cognitive dysfunction, depression, falls in the past 6 months, number of medications, hypertension, or vision/hearing problems.

Of the 26 with hypoglycemia, 12 had HbA_1c levels above 9%. “Even a high A_1c doesn't preclude lows. Hemoglobin A_1c measures the mean. There are wide fluctuations in this population,” Dr. Munshi noted.

There were a total of 102 hypoglycemic episodes, with a mean duration of 3 hours per patient. Nocturnal episodes lasted for a mean of 2.5 hours. “The duration of episodes was quite concerning,” she said.

Surprisingly, more than half (58%) of the 102 hypoglycemic episodes occurred among the 16 patients with type 2 diabetes, with a mean duration of nocturnal hypoglycemia nearly twice that of the hypoglycemia in the type 1 patients (2.9 vs. 1.6 hours). “Even the type 2 patients had wide glycemic excursions,” Dr. Munshi commented.

Also of concern, the majority of episodes (95 of the 102) were not recognized by fingerstick testing or by the patients' symptoms. Moreover, there were no significant relationships between severity of hypoglycemia and age, type of diabetes, duration, HbA_1c, treatment, or living alone.

In a follow-up interview, Dr. Munshi cited a previous study she and her colleagues published last year, showing that simplification of complex insulin regimens by using C peptide to assess whether patients could eliminate or reduce the amount of insulin taken, and adding oral agents instead, reduced hypoglycemic episodes without deterioration of glycemic control (Am. J. Med. 2009;122:395-7).

“Elderly patients with other comorbidities are unable to follow complex insulin regimens appropriately and end up having wide fluctuations in their glucose values. If a treatment regimen is designed with consideration for an elderly patient's self-care abilities, risk of hypoglycemia can be reduced,” she said.

And when available, CGM can be extremely helpful for elderly patients, Dr. Munshi said.

Major Finding: One or more hypoglycemic events occurred in 26 of 40 patients despite hemoglobin A_1c values above 8%. Of 102 hypoglycemic episodes, 95 were not recognized by fingerstick testing or by symptoms.

Data Source: Study using blinded continuous glucose monitoring for 72 or more hours in 40 community-dwelling diabetes patients seen at a tertiary diabetes center.

Disclosures: The American Diabetes Association funded the study. Dr. Munshi stated that she had no conflicts of interest.

Orlando — Hypoglycemic episodes were common despite high hemoglobin A_1c levels among 40 elderly community-living adults with diabetes who underwent continuous glucose monitoring.

“Raising A_1c goals may not be adequate to prevent hypoglycemia in this vulnerable population,” Dr. Medha N. Munshi said.

Guidelines from the American Diabetes Association and the American Geriatric Society advise that the usual recommended hemoglobin A_1c target of less than 6.5%-7.0% might be relaxed for elderly adults who have a history of severe hypoglycemia, limited life expectancy, advanced diabetes complications, or extensive comorbidity (Diabetes Care 2010;33 [suppl 1]).

In practice, this has been interpreted as a goal of less than 8%. The ADA guideline was based on level “C” evidence, and no study has ever clearly demonstrated that raising the HbA_1c target actually reduces the risk of hypoglycemia, said Dr. Munshi, director of the Joslin Diabetes Center geriatric programs at Beth Israel Deaconess Medical Center, Boston.

The current study used blinded continuous glucose monitoring (CGM) for 72 hours or longer in 40 community-dwelling diabetes patients who were seen at the Joslin Diabetes Center. To be included, they had to be older than 69 years and have a hemoglobin A_1c level greater than 8%.

The study group was 60% female and 80% white. Mean age was 75 years and mean HbA_1c was 9.3%. The patients took a mean of eight medications per day, with more than half (55%) on insulin alone, and another 38% on insulin plus one or more oral agents. Two-thirds had type 2 diabetes and the rest had type 1. Nearly a quarter were living alone, said Dr. Munshi, also with Harvard Medical School, Boston.

Patients performed four fingerstick glucose measurements per day and kept daily diaries of hypoglycemic symptoms, diet, and physical activity. One or more hypoglycemic events, defined as a glucose value less than 70 mg/dL, occurred in 26 of the 40 patients. Nearly three-fourths of the events involved glucose levels of 50-59 mg/dL, and in just under half the glucose dropped below 50 mg/dL.

The 26 patients experiencing hypoglycemia did not differ from the 14 without such events in patient characteristics including age, diabetes duration, HbA_1c, or insulin treatment. There were also no differences between those who did and did not have hypoglycemia in the comorbidities cognitive dysfunction, depression, falls in the past 6 months, number of medications, hypertension, or vision/hearing problems.

Of the 26 with hypoglycemia, 12 had HbA_1c levels above 9%. “Even a high A_1c doesn't preclude lows. Hemoglobin A_1c measures the mean. There are wide fluctuations in this population,” Dr. Munshi noted.

There were a total of 102 hypoglycemic episodes, with a mean duration of 3 hours per patient. Nocturnal episodes lasted for a mean of 2.5 hours. “The duration of episodes was quite concerning,” she said.

Surprisingly, more than half (58%) of the 102 hypoglycemic episodes occurred among the 16 patients with type 2 diabetes, with a mean duration of nocturnal hypoglycemia nearly twice that of the hypoglycemia in the type 1 patients (2.9 vs. 1.6 hours). “Even the type 2 patients had wide glycemic excursions,” Dr. Munshi commented.

Also of concern, the majority of episodes (95 of the 102) were not recognized by fingerstick testing or by the patients' symptoms. Moreover, there were no significant relationships between severity of hypoglycemia and age, type of diabetes, duration, HbA_1c, treatment, or living alone.

In a follow-up interview, Dr. Munshi cited a previous study she and her colleagues published last year, showing that simplification of complex insulin regimens by using C peptide to assess whether patients could eliminate or reduce the amount of insulin taken, and adding oral agents instead, reduced hypoglycemic episodes without deterioration of glycemic control (Am. J. Med. 2009;122:395-7).

“Elderly patients with other comorbidities are unable to follow complex insulin regimens appropriately and end up having wide fluctuations in their glucose values. If a treatment regimen is designed with consideration for an elderly patient's self-care abilities, risk of hypoglycemia can be reduced,” she said.

And when available, CGM can be extremely helpful for elderly patients, Dr. Munshi said.

ORLANDO — Hip and distal fractures were found to be elevated both in women and men who had taken thiazolidinediones in an analysis of all hospitalizations for fracture in Scotland during 2000-2008.

The current product labels for rosiglitazone and pioglitazone state that the fracture risk associated with the thiazolidinedione (TZD) class is predominantly for distal fractures in women. But the current study, which included an analysis using patients as their own controls to overcome treatment allocation bias, showed that the risk of hip fractures is substantially elevated in men and women. The findings suggest that the Food and Drug Administration should consider whether a change in drug label is indicated, Dr. Helen Colhoun, professor of public health at the University of Dundee, Scotland, said at the meeting. No difference in risk was noted between the two individual TZDs, she noted.

A clinical database identified all individuals in Scotland aged at least 40 with diagnosed type 2 diabetes, and linked those data with hospital admission records. The study comprised a total 82,027 person-years of TZD use in 35,606 users (68% rosiglitazone, 32% pioglitazone) and 6,209 fractures in 5,094 individuals.

There were 1,375 fractures among the 35,606 with diabetes who had used TZDs (3.9%) and 4,835 among the 112,100 (4.3%) who had never used TZDs. Patients who had used TZDs were significantly younger, had a higher body mass index, were less likely to have had a prior fracture, and less likely to have had recent cardiovascular disease.

After adjustment for diabetes duration, prior CVD, baseline and time-dependent exposure to other medications, and BMI, the rate of fracture per cumulative year of TZD exposure was significantly elevated for hip fracture in both men (hazard ratio, 1.14) and women (HR, 1.17) and for distal fracture in women (1.16).

In the self-controlled case series, the incidence rate ratio comparing the exposed and unexposed periods of time for each individual patient demonstrated a significantly increased risk of hip fractures during the period of TZD exposure for the group as a whole (incidence rate ratio, 1.98) and in both men (2.23) and women (1.9). The risk for distal fractures was significantly greater for the entire group (1.4) and for women (1.57) but was not significant among men.

The incidence rate ratio by duration of TZD exposure rose from 1.45 at less than 1 year to 3.59 after 6 years of exposure, Dr. Colhoun reported.

The overall relative risk in the self-controlled case series of approximately 2 and a current prevalence of exposure of about 20% implies that about 17% of fractures in patients with diabetes could be attributable to TZDs, she noted.

Of note, she added, this analysis assumes that the absence of TZDs would not result in more use of another agent that also increases fracture risk such as insulin. Studies are underway to assess that further and to build a predictive model of fracture in diabetes. Other studies are looking at whether reducing the TZD dose might lower the fracture risk, and whether the risk goes away when the drug is stopped. Preliminary data suggest that the risk remains high, she said.

Wellcome Trust Scottish Health Informatics Programme and the Chief Scientist's Office, Scotland, funded the study. Dr. Colhoun disclosed that she has received research support from Pfizer, Eli Lilly, Boehringer Ingelheim, AstraZeneca, and Roche. She serves on the speakers bureau for Pfizer, is a board/advisory panel member for Eli Lilly, and owns stocks/shares in Roche.

ORLANDO — Hip and distal fractures were found to be elevated both in women and men who had taken thiazolidinediones in an analysis of all hospitalizations for fracture in Scotland during 2000-2008.

The current product labels for rosiglitazone and pioglitazone state that the fracture risk associated with the thiazolidinedione (TZD) class is predominantly for distal fractures in women. But the current study, which included an analysis using patients as their own controls to overcome treatment allocation bias, showed that the risk of hip fractures is substantially elevated in men and women. The findings suggest that the Food and Drug Administration should consider whether a change in drug label is indicated, Dr. Helen Colhoun, professor of public health at the University of Dundee, Scotland, said at the meeting. No difference in risk was noted between the two individual TZDs, she noted.

A clinical database identified all individuals in Scotland aged at least 40 with diagnosed type 2 diabetes, and linked those data with hospital admission records. The study comprised a total 82,027 person-years of TZD use in 35,606 users (68% rosiglitazone, 32% pioglitazone) and 6,209 fractures in 5,094 individuals.

There were 1,375 fractures among the 35,606 with diabetes who had used TZDs (3.9%) and 4,835 among the 112,100 (4.3%) who had never used TZDs. Patients who had used TZDs were significantly younger, had a higher body mass index, were less likely to have had a prior fracture, and less likely to have had recent cardiovascular disease.

After adjustment for diabetes duration, prior CVD, baseline and time-dependent exposure to other medications, and BMI, the rate of fracture per cumulative year of TZD exposure was significantly elevated for hip fracture in both men (hazard ratio, 1.14) and women (HR, 1.17) and for distal fracture in women (1.16).

In the self-controlled case series, the incidence rate ratio comparing the exposed and unexposed periods of time for each individual patient demonstrated a significantly increased risk of hip fractures during the period of TZD exposure for the group as a whole (incidence rate ratio, 1.98) and in both men (2.23) and women (1.9). The risk for distal fractures was significantly greater for the entire group (1.4) and for women (1.57) but was not significant among men.

The incidence rate ratio by duration of TZD exposure rose from 1.45 at less than 1 year to 3.59 after 6 years of exposure, Dr. Colhoun reported.

The overall relative risk in the self-controlled case series of approximately 2 and a current prevalence of exposure of about 20% implies that about 17% of fractures in patients with diabetes could be attributable to TZDs, she noted.

Of note, she added, this analysis assumes that the absence of TZDs would not result in more use of another agent that also increases fracture risk such as insulin. Studies are underway to assess that further and to build a predictive model of fracture in diabetes. Other studies are looking at whether reducing the TZD dose might lower the fracture risk, and whether the risk goes away when the drug is stopped. Preliminary data suggest that the risk remains high, she said.

Wellcome Trust Scottish Health Informatics Programme and the Chief Scientist's Office, Scotland, funded the study. Dr. Colhoun disclosed that she has received research support from Pfizer, Eli Lilly, Boehringer Ingelheim, AstraZeneca, and Roche. She serves on the speakers bureau for Pfizer, is a board/advisory panel member for Eli Lilly, and owns stocks/shares in Roche.

ORLANDO — Hip and distal fractures were found to be elevated both in women and men who had taken thiazolidinediones in an analysis of all hospitalizations for fracture in Scotland during 2000-2008.

The current product labels for rosiglitazone and pioglitazone state that the fracture risk associated with the thiazolidinedione (TZD) class is predominantly for distal fractures in women. But the current study, which included an analysis using patients as their own controls to overcome treatment allocation bias, showed that the risk of hip fractures is substantially elevated in men and women. The findings suggest that the Food and Drug Administration should consider whether a change in drug label is indicated, Dr. Helen Colhoun, professor of public health at the University of Dundee, Scotland, said at the meeting. No difference in risk was noted between the two individual TZDs, she noted.

A clinical database identified all individuals in Scotland aged at least 40 with diagnosed type 2 diabetes, and linked those data with hospital admission records. The study comprised a total 82,027 person-years of TZD use in 35,606 users (68% rosiglitazone, 32% pioglitazone) and 6,209 fractures in 5,094 individuals.

There were 1,375 fractures among the 35,606 with diabetes who had used TZDs (3.9%) and 4,835 among the 112,100 (4.3%) who had never used TZDs. Patients who had used TZDs were significantly younger, had a higher body mass index, were less likely to have had a prior fracture, and less likely to have had recent cardiovascular disease.

After adjustment for diabetes duration, prior CVD, baseline and time-dependent exposure to other medications, and BMI, the rate of fracture per cumulative year of TZD exposure was significantly elevated for hip fracture in both men (hazard ratio, 1.14) and women (HR, 1.17) and for distal fracture in women (1.16).

In the self-controlled case series, the incidence rate ratio comparing the exposed and unexposed periods of time for each individual patient demonstrated a significantly increased risk of hip fractures during the period of TZD exposure for the group as a whole (incidence rate ratio, 1.98) and in both men (2.23) and women (1.9). The risk for distal fractures was significantly greater for the entire group (1.4) and for women (1.57) but was not significant among men.

The incidence rate ratio by duration of TZD exposure rose from 1.45 at less than 1 year to 3.59 after 6 years of exposure, Dr. Colhoun reported.

The overall relative risk in the self-controlled case series of approximately 2 and a current prevalence of exposure of about 20% implies that about 17% of fractures in patients with diabetes could be attributable to TZDs, she noted.

Of note, she added, this analysis assumes that the absence of TZDs would not result in more use of another agent that also increases fracture risk such as insulin. Studies are underway to assess that further and to build a predictive model of fracture in diabetes. Other studies are looking at whether reducing the TZD dose might lower the fracture risk, and whether the risk goes away when the drug is stopped. Preliminary data suggest that the risk remains high, she said.

Wellcome Trust Scottish Health Informatics Programme and the Chief Scientist's Office, Scotland, funded the study. Dr. Colhoun disclosed that she has received research support from Pfizer, Eli Lilly, Boehringer Ingelheim, AstraZeneca, and Roche. She serves on the speakers bureau for Pfizer, is a board/advisory panel member for Eli Lilly, and owns stocks/shares in Roche.

SAN ANTONIO – Data continue to show that diabetes education saves money.

Last year, the findings of an analysis of 3 years’ worth of data from a large national claims database showed that patients with diabetes who receive diabetes education incur lower costs than do those who have not received the education (Diabetes Educ. 2009;35:752-60).

Now, preliminary findings from a new analysis of the same database show that the cost reduction comes primarily from a drop in hospital admissions, and that ongoing diabetes education beyond the initial sessions received at diagnosis are necessary to produce the cost savings.

The findings were presented by actuarial consultant Ian Duncan, president of Solucia Consulting. His firm conducted the analysis for the American Association of Diabetes Educators using data from Solucia’s claims database of about 20 million individuals, including both Medicare and commercially insured members. In 2007, the database contained 166,931 individuals with diabetes who had commercial insurance and 56,345 who were on Medicare.

“We all know how much you do for patients, how much you improve their lifestyle and help them cope with their condition. But the kinds of people we work with – the insurance payers (Medicaid agencies, Medicare, and health plans) – are interested in whether the investment for the services you provide brings a return to them in terms of the costs and the claims that the patients incur,” said Mr. Duncan, also of the department of health administration at Georgetown University in Washington.

In 2007, the adjusted cost per member per month of plan members with diabetes who were commercially insured was $923 among those who received diabetes education, compared with $1,072 among those who did not. For Medicare enrollees, those figures were $1,241 and $1,322, respectively. Those differences were significant for both groups.

The follow-up, risk-adjusted analysis of patients who were continuously enrolled in 2005-2008 showed a significant difference in costs for the commercially insured patients ($985 for the 3,094 who received diabetes education vs. $1,043 for the 31,075 who did not), but there were no differences in the Medicare group, which included just 898 patients who received diabetes education and 23,342 who did not (nearly $1,400 for both).

However, when the frequency of diabetes education was factored in, it became clear that enrollees who received two or more sessions of diabetes education incurred lower costs than did those who received just one or no sessions.

In the commercially insured group in 2008, costs per patient per month were $845 for those receiving at least two education sessions, $863 for those who had just one session, and $907 for those with no diabetes education. In the Medicare group, the costs for those with zero and one session were nearly identical ($1,343 and $1,337, respectively), whereas the patients who did receive diabetes education that was covered by Medicare cost $1,267.50 per month. Neither quite reached statistical significance.

“This year’s data show that follow-up diabetes education is what really produces the outstanding results,” Mr. Duncan commented, to audience applause.

These differences were seen even though more diabetes education was associated with higher medication compliance and thus greater pharmacy costs. In 2008, costs for glucose-lowering drugs were nearly identical for those with commercial insurance who received no education sessions or just one ($76 and $78, respectively) but were significantly higher for those who had at least two sessions ($99). Similarly, those figures in the Medicare group were $69 and $70, compared with $81, respectively.

Reductions in hospital admissions in the group receiving diabetes education more than made up for the higher pharmacy cost. In 2008, there were just 180 per 1,000 admissions for diabetes patients with commercial insurance who received two or more diabetes education sessions, compared with 212 per 1,000 for those with one session and 221 per 1,000 for no sessions. The difference was not as striking in Medicare, where those figures were 709, 665, and 735 per 1,000, respectively.

This study is ongoing, with additional analyses of laboratory and clinical values as well as measures of behavior change, he said.

Other than the AADE funding for this study, Mr. Duncan stated that he had no further disclosures.

SAN ANTONIO – Data continue to show that diabetes education saves money.

Last year, the findings of an analysis of 3 years’ worth of data from a large national claims database showed that patients with diabetes who receive diabetes education incur lower costs than do those who have not received the education (Diabetes Educ. 2009;35:752-60).

Now, preliminary findings from a new analysis of the same database show that the cost reduction comes primarily from a drop in hospital admissions, and that ongoing diabetes education beyond the initial sessions received at diagnosis are necessary to produce the cost savings.

The findings were presented by actuarial consultant Ian Duncan, president of Solucia Consulting. His firm conducted the analysis for the American Association of Diabetes Educators using data from Solucia’s claims database of about 20 million individuals, including both Medicare and commercially insured members. In 2007, the database contained 166,931 individuals with diabetes who had commercial insurance and 56,345 who were on Medicare.

“We all know how much you do for patients, how much you improve their lifestyle and help them cope with their condition. But the kinds of people we work with – the insurance payers (Medicaid agencies, Medicare, and health plans) – are interested in whether the investment for the services you provide brings a return to them in terms of the costs and the claims that the patients incur,” said Mr. Duncan, also of the department of health administration at Georgetown University in Washington.

In 2007, the adjusted cost per member per month of plan members with diabetes who were commercially insured was $923 among those who received diabetes education, compared with $1,072 among those who did not. For Medicare enrollees, those figures were $1,241 and $1,322, respectively. Those differences were significant for both groups.

The follow-up, risk-adjusted analysis of patients who were continuously enrolled in 2005-2008 showed a significant difference in costs for the commercially insured patients ($985 for the 3,094 who received diabetes education vs. $1,043 for the 31,075 who did not), but there were no differences in the Medicare group, which included just 898 patients who received diabetes education and 23,342 who did not (nearly $1,400 for both).

However, when the frequency of diabetes education was factored in, it became clear that enrollees who received two or more sessions of diabetes education incurred lower costs than did those who received just one or no sessions.

In the commercially insured group in 2008, costs per patient per month were $845 for those receiving at least two education sessions, $863 for those who had just one session, and $907 for those with no diabetes education. In the Medicare group, the costs for those with zero and one session were nearly identical ($1,343 and $1,337, respectively), whereas the patients who did receive diabetes education that was covered by Medicare cost $1,267.50 per month. Neither quite reached statistical significance.

“This year’s data show that follow-up diabetes education is what really produces the outstanding results,” Mr. Duncan commented, to audience applause.

These differences were seen even though more diabetes education was associated with higher medication compliance and thus greater pharmacy costs. In 2008, costs for glucose-lowering drugs were nearly identical for those with commercial insurance who received no education sessions or just one ($76 and $78, respectively) but were significantly higher for those who had at least two sessions ($99). Similarly, those figures in the Medicare group were $69 and $70, compared with $81, respectively.

Reductions in hospital admissions in the group receiving diabetes education more than made up for the higher pharmacy cost. In 2008, there were just 180 per 1,000 admissions for diabetes patients with commercial insurance who received two or more diabetes education sessions, compared with 212 per 1,000 for those with one session and 221 per 1,000 for no sessions. The difference was not as striking in Medicare, where those figures were 709, 665, and 735 per 1,000, respectively.

This study is ongoing, with additional analyses of laboratory and clinical values as well as measures of behavior change, he said.

Other than the AADE funding for this study, Mr. Duncan stated that he had no further disclosures.

SAN ANTONIO – Data continue to show that diabetes education saves money.

Last year, the findings of an analysis of 3 years’ worth of data from a large national claims database showed that patients with diabetes who receive diabetes education incur lower costs than do those who have not received the education (Diabetes Educ. 2009;35:752-60).

Now, preliminary findings from a new analysis of the same database show that the cost reduction comes primarily from a drop in hospital admissions, and that ongoing diabetes education beyond the initial sessions received at diagnosis are necessary to produce the cost savings.

The findings were presented by actuarial consultant Ian Duncan, president of Solucia Consulting. His firm conducted the analysis for the American Association of Diabetes Educators using data from Solucia’s claims database of about 20 million individuals, including both Medicare and commercially insured members. In 2007, the database contained 166,931 individuals with diabetes who had commercial insurance and 56,345 who were on Medicare.

“We all know how much you do for patients, how much you improve their lifestyle and help them cope with their condition. But the kinds of people we work with – the insurance payers (Medicaid agencies, Medicare, and health plans) – are interested in whether the investment for the services you provide brings a return to them in terms of the costs and the claims that the patients incur,” said Mr. Duncan, also of the department of health administration at Georgetown University in Washington.

In 2007, the adjusted cost per member per month of plan members with diabetes who were commercially insured was $923 among those who received diabetes education, compared with $1,072 among those who did not. For Medicare enrollees, those figures were $1,241 and $1,322, respectively. Those differences were significant for both groups.

The follow-up, risk-adjusted analysis of patients who were continuously enrolled in 2005-2008 showed a significant difference in costs for the commercially insured patients ($985 for the 3,094 who received diabetes education vs. $1,043 for the 31,075 who did not), but there were no differences in the Medicare group, which included just 898 patients who received diabetes education and 23,342 who did not (nearly $1,400 for both).

However, when the frequency of diabetes education was factored in, it became clear that enrollees who received two or more sessions of diabetes education incurred lower costs than did those who received just one or no sessions.

In the commercially insured group in 2008, costs per patient per month were $845 for those receiving at least two education sessions, $863 for those who had just one session, and $907 for those with no diabetes education. In the Medicare group, the costs for those with zero and one session were nearly identical ($1,343 and $1,337, respectively), whereas the patients who did receive diabetes education that was covered by Medicare cost $1,267.50 per month. Neither quite reached statistical significance.

“This year’s data show that follow-up diabetes education is what really produces the outstanding results,” Mr. Duncan commented, to audience applause.

These differences were seen even though more diabetes education was associated with higher medication compliance and thus greater pharmacy costs. In 2008, costs for glucose-lowering drugs were nearly identical for those with commercial insurance who received no education sessions or just one ($76 and $78, respectively) but were significantly higher for those who had at least two sessions ($99). Similarly, those figures in the Medicare group were $69 and $70, compared with $81, respectively.

Reductions in hospital admissions in the group receiving diabetes education more than made up for the higher pharmacy cost. In 2008, there were just 180 per 1,000 admissions for diabetes patients with commercial insurance who received two or more diabetes education sessions, compared with 212 per 1,000 for those with one session and 221 per 1,000 for no sessions. The difference was not as striking in Medicare, where those figures were 709, 665, and 735 per 1,000, respectively.

This study is ongoing, with additional analyses of laboratory and clinical values as well as measures of behavior change, he said.

Other than the AADE funding for this study, Mr. Duncan stated that he had no further disclosures.

SAN ANTONIO  – Overweight individuals who knew that they had prediabetes did not lose any more weight than did those who did not have the condition, in a preliminary analysis of 103 overweight participants in a 12-week behavioral weight-loss program.

The findings were contrary to expectations. “Our hypothesis was that individuals who perceive themselves to be at increased risk for developing diabetes would lose more weight than individuals who perceive their risk to be lower,” said Carol J. Homko, Ph.D., a certified diabetes educator and registered nurse.

The data come from a secondary analysis of a larger randomized study that was designed to look at the effect of telemedicine on weight-loss maintenance 1 year following the weight-loss program. All 103 subjects participated in the same weekly group sessions, which were held at local churches and facilitated by both health care professionals (mostly registered dieticians and health psychologists) and a lay facilitator who was a member of the church. The 90-minute sessions included diet, exercise, and behavioral change counseling. Each group included about 12-15 participants.

The participants had a mean age of 51 years (adults aged 18-75 were included) and body mass index of 35.5 kg/m2. Nearly all were female (96%) and black (97%). At baseline, oral glucose tolerance testing revealed that 53 had impaired glucose tolerance (“prediabetes”), whereas the other 50 were normoglycemic. Those with overt diabetes were excluded. Individuals with and without prediabetes did not differ at baseline in weight, BMI, systolic blood pressure, or cholesterol level.

Risk perception was assessed using the same survey used in the Diabetes Prevention Program (Diabetes Care 2003;26:2543-8), and included measures of personal control, worry, optimistic bias, comparative disease risk, comparative environmental risk, diabetes risk knowledge, unknown risk, and “dread risk.”

Across the board, there were no differences in risk perception from baseline through 12 weeks in any of those measures, said Dr. Homko of the department of medicine at Temple University, Philadelphia.

It’s not clear why people who know they have prediabetes don’t perceive themselves to be at greater risk for diabetes than are those without that diagnosis, but one possible explanation is that the people who chose to participate in a weight-loss program did so because they were already aware of the risks associated with excess weight and the potential impact on their health, she said in an interview.

Participants lost a mean of 6.3 pounds, and weight losses did not differ significantly between those with and without prediabetes (7.0 vs. 5.6 pounds, respectively).

Not surprisingly, weight loss correlated with progression to diabetes among the 53 with prediabetes. The 28 who reverted to normal glucose tolerance had lost an average of 11.3 pounds at 12 weeks, compared with 5.6 pounds in the 18 who remained prediabetic and 3.6 pounds in the 7 who progressed to type 2 diabetes. Those differences were statistically significant, she said.

Although the findings from this study regarding risk perception may be disappointing, the bright side is that weight loss of just 5% was associated with conversion to normal glucose tolerance in individuals with prediabetes, Dr. Homko pointed out. “There is strong evidence to suggest that weight loss and increased activity can slow the progression to type 2 diabetes,” she said.

Dr. Homko disclosed that she serves on the advisory board of Abbott Diabetes Care.

SAN ANTONIO  – Overweight individuals who knew that they had prediabetes did not lose any more weight than did those who did not have the condition, in a preliminary analysis of 103 overweight participants in a 12-week behavioral weight-loss program.

The findings were contrary to expectations. “Our hypothesis was that individuals who perceive themselves to be at increased risk for developing diabetes would lose more weight than individuals who perceive their risk to be lower,” said Carol J. Homko, Ph.D., a certified diabetes educator and registered nurse.

The data come from a secondary analysis of a larger randomized study that was designed to look at the effect of telemedicine on weight-loss maintenance 1 year following the weight-loss program. All 103 subjects participated in the same weekly group sessions, which were held at local churches and facilitated by both health care professionals (mostly registered dieticians and health psychologists) and a lay facilitator who was a member of the church. The 90-minute sessions included diet, exercise, and behavioral change counseling. Each group included about 12-15 participants.

The participants had a mean age of 51 years (adults aged 18-75 were included) and body mass index of 35.5 kg/m2. Nearly all were female (96%) and black (97%). At baseline, oral glucose tolerance testing revealed that 53 had impaired glucose tolerance (“prediabetes”), whereas the other 50 were normoglycemic. Those with overt diabetes were excluded. Individuals with and without prediabetes did not differ at baseline in weight, BMI, systolic blood pressure, or cholesterol level.

Risk perception was assessed using the same survey used in the Diabetes Prevention Program (Diabetes Care 2003;26:2543-8), and included measures of personal control, worry, optimistic bias, comparative disease risk, comparative environmental risk, diabetes risk knowledge, unknown risk, and “dread risk.”

Across the board, there were no differences in risk perception from baseline through 12 weeks in any of those measures, said Dr. Homko of the department of medicine at Temple University, Philadelphia.

It’s not clear why people who know they have prediabetes don’t perceive themselves to be at greater risk for diabetes than are those without that diagnosis, but one possible explanation is that the people who chose to participate in a weight-loss program did so because they were already aware of the risks associated with excess weight and the potential impact on their health, she said in an interview.

Participants lost a mean of 6.3 pounds, and weight losses did not differ significantly between those with and without prediabetes (7.0 vs. 5.6 pounds, respectively).

Not surprisingly, weight loss correlated with progression to diabetes among the 53 with prediabetes. The 28 who reverted to normal glucose tolerance had lost an average of 11.3 pounds at 12 weeks, compared with 5.6 pounds in the 18 who remained prediabetic and 3.6 pounds in the 7 who progressed to type 2 diabetes. Those differences were statistically significant, she said.

Although the findings from this study regarding risk perception may be disappointing, the bright side is that weight loss of just 5% was associated with conversion to normal glucose tolerance in individuals with prediabetes, Dr. Homko pointed out. “There is strong evidence to suggest that weight loss and increased activity can slow the progression to type 2 diabetes,” she said.

Dr. Homko disclosed that she serves on the advisory board of Abbott Diabetes Care.

SAN ANTONIO  – Overweight individuals who knew that they had prediabetes did not lose any more weight than did those who did not have the condition, in a preliminary analysis of 103 overweight participants in a 12-week behavioral weight-loss program.

The findings were contrary to expectations. “Our hypothesis was that individuals who perceive themselves to be at increased risk for developing diabetes would lose more weight than individuals who perceive their risk to be lower,” said Carol J. Homko, Ph.D., a certified diabetes educator and registered nurse.

The data come from a secondary analysis of a larger randomized study that was designed to look at the effect of telemedicine on weight-loss maintenance 1 year following the weight-loss program. All 103 subjects participated in the same weekly group sessions, which were held at local churches and facilitated by both health care professionals (mostly registered dieticians and health psychologists) and a lay facilitator who was a member of the church. The 90-minute sessions included diet, exercise, and behavioral change counseling. Each group included about 12-15 participants.

The participants had a mean age of 51 years (adults aged 18-75 were included) and body mass index of 35.5 kg/m2. Nearly all were female (96%) and black (97%). At baseline, oral glucose tolerance testing revealed that 53 had impaired glucose tolerance (“prediabetes”), whereas the other 50 were normoglycemic. Those with overt diabetes were excluded. Individuals with and without prediabetes did not differ at baseline in weight, BMI, systolic blood pressure, or cholesterol level.

Risk perception was assessed using the same survey used in the Diabetes Prevention Program (Diabetes Care 2003;26:2543-8), and included measures of personal control, worry, optimistic bias, comparative disease risk, comparative environmental risk, diabetes risk knowledge, unknown risk, and “dread risk.”

Across the board, there were no differences in risk perception from baseline through 12 weeks in any of those measures, said Dr. Homko of the department of medicine at Temple University, Philadelphia.

It’s not clear why people who know they have prediabetes don’t perceive themselves to be at greater risk for diabetes than are those without that diagnosis, but one possible explanation is that the people who chose to participate in a weight-loss program did so because they were already aware of the risks associated with excess weight and the potential impact on their health, she said in an interview.

Participants lost a mean of 6.3 pounds, and weight losses did not differ significantly between those with and without prediabetes (7.0 vs. 5.6 pounds, respectively).

Not surprisingly, weight loss correlated with progression to diabetes among the 53 with prediabetes. The 28 who reverted to normal glucose tolerance had lost an average of 11.3 pounds at 12 weeks, compared with 5.6 pounds in the 18 who remained prediabetic and 3.6 pounds in the 7 who progressed to type 2 diabetes. Those differences were statistically significant, she said.

Although the findings from this study regarding risk perception may be disappointing, the bright side is that weight loss of just 5% was associated with conversion to normal glucose tolerance in individuals with prediabetes, Dr. Homko pointed out. “There is strong evidence to suggest that weight loss and increased activity can slow the progression to type 2 diabetes,” she said.

Dr. Homko disclosed that she serves on the advisory board of Abbott Diabetes Care.

SAN ANTONIO – A personalized phone call from a retail chain store pharmacist to patients who missed diabetes prescription refills significantly improved medication adherence at 1 year in a study of 265 patients with type 2 diabetes.

“Because adults with diabetes visit pharmacies more often than they visit any other health professional [setting], it is believed there is an untapped opportunity for pharmacists to provide self-management education and support for medications,” said certified diabetes educator Peggy S. Odegard, Pharm.D., at the annual meeting of the American Association of Diabetes Educators .

The randomized, controlled Medication Adherence Program (MAP) study was conducted at four pharmacies inside Safeway grocery stores in Washington state. When a prescription refill for an oral glucose-lowering medication was missed by 6 days, a pharmacist would call the patients to ask why they had missed the refill and whether they would like to refill it now. Depending on the response, the pharmacist would offer individualized advice and education. A follow-up phone call was made at 1 week to 1 month after the intervention to further assess the patients’ needs and address any problems.

The subjects had all been using the pharmacy consistently for a year or longer. The 145 controls, who were not called when refills were missed, were slightly younger, with a mean age of 61 years, compared with 65 years for the 120 who received the phone call intervention. The groups’ other baseline characteristics were similar, including sex (a little more than half were women), the number of different medications they were taking, and the proportion who were on insulin (23% in both groups).

Among 119 patients who reported problems with taking their medications, 27% cited “difficulty with taking medication,” 26% said they simply “forgot to order refills,” and 8% “forgot to pick up refills.” Of those with “difficulty taking medication,” the most common difficulty listed was “remembering dose.”

Adherence was assessed by the change in “medication possession ratio (MPR),” or the number of days of medication supplied in a prescription fill divided by the number of days until the prescription was refilled. For example, a 30-day supply that is filled and then refilled in 30 days would yield an MPR of 1.0. But if a patient receives a 30-day prescription but doesn’t refill it for 60 days, the MPR would be 0.5, or half the adherence rate expected, explained Dr. Odegard, of the University of Washington, Seattle, in an interview.

At baseline, MPR for diabetes medications did not differ between the two groups (0.86 for the intervention group and 0.84 for controls). However, the proportion of patients with an MPR greater than 0.8 was slightly higher in the intervention group than in the controls (74.4% vs. 65.2%), meaning that it would be harder to prove that the intervention worked because the patients in that group already were somewhat more adherent, Dr. Odegard pointed out.

Over 12 months, the pharmacists conducted an average of 3.4 phone call interventions (or occasionally in-person interventions) per patient, and were reimbursed $10 per intervention. Interventions lasted an average of 12.6 minutes each. In addition to diabetes education (including advice on prevention of medication side effects) and adherence support (including integration of medication dosing with daily activities), pharmacists helped to optimize the patients’ regimen with the prescriber and/or helped with economic adjustment (for example, a change to generic).

At 12 months, the MPR was significantly improved in the intervention group (up to 0.90 from 0.86), whereas in the control group the MPR declined slightly (from 0.84 to 0.82). The difference in MPR between the two groups at 12 months also was significant. Moreover, the likelihood of an oral medication adherence rate greater than 80% (MPR 0.80 or higher) was 4.77 times greater among the intervention group than in the controls. This difference was significant despite wide confidence intervals, said Dr. Odegard.

A regression model that included demographic variables, regimen complexity, and prior oral medication adherence measures explained 70% of the change in MPR scores at 12 months. Significant predictors of a change in MPR were a low MPR during the prior period, less resistance to taking medication, and less time needed with the pharmacist.

Dr. Odegard and her associates are hoping to integrate this type of program into local pharmacy chains. Several remaining challenges include the fact that patients don’t always pick up their own medications, some might use both community and mail-order pharmacies, and some might have multiple physicians for their diabetes care.

During the question-and-answer period, Dr. Odegard remarked that such programs could provide a significant revenue stream to pharmacies. She and her colleagues are currently conducting a cost analysis.

Indeed, Safeway has been very supportive. “If a company like Safeway can take the lead on helping drive corporate policy change in helping pharmacists to do this kind of thing, if it really makes a difference, it will be huge. ... We’re really hoping that reimbursement can go to pharmacies or other health care providers who perform these services. Corporate policy change is really important to help our patients get what they need,” she said.

SAN ANTONIO – A personalized phone call from a retail chain store pharmacist to patients who missed diabetes prescription refills significantly improved medication adherence at 1 year in a study of 265 patients with type 2 diabetes.

“Because adults with diabetes visit pharmacies more often than they visit any other health professional [setting], it is believed there is an untapped opportunity for pharmacists to provide self-management education and support for medications,” said certified diabetes educator Peggy S. Odegard, Pharm.D., at the annual meeting of the American Association of Diabetes Educators .

The randomized, controlled Medication Adherence Program (MAP) study was conducted at four pharmacies inside Safeway grocery stores in Washington state. When a prescription refill for an oral glucose-lowering medication was missed by 6 days, a pharmacist would call the patients to ask why they had missed the refill and whether they would like to refill it now. Depending on the response, the pharmacist would offer individualized advice and education. A follow-up phone call was made at 1 week to 1 month after the intervention to further assess the patients’ needs and address any problems.

The subjects had all been using the pharmacy consistently for a year or longer. The 145 controls, who were not called when refills were missed, were slightly younger, with a mean age of 61 years, compared with 65 years for the 120 who received the phone call intervention. The groups’ other baseline characteristics were similar, including sex (a little more than half were women), the number of different medications they were taking, and the proportion who were on insulin (23% in both groups).

Among 119 patients who reported problems with taking their medications, 27% cited “difficulty with taking medication,” 26% said they simply “forgot to order refills,” and 8% “forgot to pick up refills.” Of those with “difficulty taking medication,” the most common difficulty listed was “remembering dose.”

Adherence was assessed by the change in “medication possession ratio (MPR),” or the number of days of medication supplied in a prescription fill divided by the number of days until the prescription was refilled. For example, a 30-day supply that is filled and then refilled in 30 days would yield an MPR of 1.0. But if a patient receives a 30-day prescription but doesn’t refill it for 60 days, the MPR would be 0.5, or half the adherence rate expected, explained Dr. Odegard, of the University of Washington, Seattle, in an interview.

At baseline, MPR for diabetes medications did not differ between the two groups (0.86 for the intervention group and 0.84 for controls). However, the proportion of patients with an MPR greater than 0.8 was slightly higher in the intervention group than in the controls (74.4% vs. 65.2%), meaning that it would be harder to prove that the intervention worked because the patients in that group already were somewhat more adherent, Dr. Odegard pointed out.

Over 12 months, the pharmacists conducted an average of 3.4 phone call interventions (or occasionally in-person interventions) per patient, and were reimbursed $10 per intervention. Interventions lasted an average of 12.6 minutes each. In addition to diabetes education (including advice on prevention of medication side effects) and adherence support (including integration of medication dosing with daily activities), pharmacists helped to optimize the patients’ regimen with the prescriber and/or helped with economic adjustment (for example, a change to generic).

At 12 months, the MPR was significantly improved in the intervention group (up to 0.90 from 0.86), whereas in the control group the MPR declined slightly (from 0.84 to 0.82). The difference in MPR between the two groups at 12 months also was significant. Moreover, the likelihood of an oral medication adherence rate greater than 80% (MPR 0.80 or higher) was 4.77 times greater among the intervention group than in the controls. This difference was significant despite wide confidence intervals, said Dr. Odegard.

A regression model that included demographic variables, regimen complexity, and prior oral medication adherence measures explained 70% of the change in MPR scores at 12 months. Significant predictors of a change in MPR were a low MPR during the prior period, less resistance to taking medication, and less time needed with the pharmacist.

Dr. Odegard and her associates are hoping to integrate this type of program into local pharmacy chains. Several remaining challenges include the fact that patients don’t always pick up their own medications, some might use both community and mail-order pharmacies, and some might have multiple physicians for their diabetes care.

During the question-and-answer period, Dr. Odegard remarked that such programs could provide a significant revenue stream to pharmacies. She and her colleagues are currently conducting a cost analysis.

Indeed, Safeway has been very supportive. “If a company like Safeway can take the lead on helping drive corporate policy change in helping pharmacists to do this kind of thing, if it really makes a difference, it will be huge. ... We’re really hoping that reimbursement can go to pharmacies or other health care providers who perform these services. Corporate policy change is really important to help our patients get what they need,” she said.

SAN ANTONIO – A personalized phone call from a retail chain store pharmacist to patients who missed diabetes prescription refills significantly improved medication adherence at 1 year in a study of 265 patients with type 2 diabetes.

“Because adults with diabetes visit pharmacies more often than they visit any other health professional [setting], it is believed there is an untapped opportunity for pharmacists to provide self-management education and support for medications,” said certified diabetes educator Peggy S. Odegard, Pharm.D., at the annual meeting of the American Association of Diabetes Educators .

The randomized, controlled Medication Adherence Program (MAP) study was conducted at four pharmacies inside Safeway grocery stores in Washington state. When a prescription refill for an oral glucose-lowering medication was missed by 6 days, a pharmacist would call the patients to ask why they had missed the refill and whether they would like to refill it now. Depending on the response, the pharmacist would offer individualized advice and education. A follow-up phone call was made at 1 week to 1 month after the intervention to further assess the patients’ needs and address any problems.

The subjects had all been using the pharmacy consistently for a year or longer. The 145 controls, who were not called when refills were missed, were slightly younger, with a mean age of 61 years, compared with 65 years for the 120 who received the phone call intervention. The groups’ other baseline characteristics were similar, including sex (a little more than half were women), the number of different medications they were taking, and the proportion who were on insulin (23% in both groups).

Among 119 patients who reported problems with taking their medications, 27% cited “difficulty with taking medication,” 26% said they simply “forgot to order refills,” and 8% “forgot to pick up refills.” Of those with “difficulty taking medication,” the most common difficulty listed was “remembering dose.”

Adherence was assessed by the change in “medication possession ratio (MPR),” or the number of days of medication supplied in a prescription fill divided by the number of days until the prescription was refilled. For example, a 30-day supply that is filled and then refilled in 30 days would yield an MPR of 1.0. But if a patient receives a 30-day prescription but doesn’t refill it for 60 days, the MPR would be 0.5, or half the adherence rate expected, explained Dr. Odegard, of the University of Washington, Seattle, in an interview.

At baseline, MPR for diabetes medications did not differ between the two groups (0.86 for the intervention group and 0.84 for controls). However, the proportion of patients with an MPR greater than 0.8 was slightly higher in the intervention group than in the controls (74.4% vs. 65.2%), meaning that it would be harder to prove that the intervention worked because the patients in that group already were somewhat more adherent, Dr. Odegard pointed out.

Over 12 months, the pharmacists conducted an average of 3.4 phone call interventions (or occasionally in-person interventions) per patient, and were reimbursed $10 per intervention. Interventions lasted an average of 12.6 minutes each. In addition to diabetes education (including advice on prevention of medication side effects) and adherence support (including integration of medication dosing with daily activities), pharmacists helped to optimize the patients’ regimen with the prescriber and/or helped with economic adjustment (for example, a change to generic).

At 12 months, the MPR was significantly improved in the intervention group (up to 0.90 from 0.86), whereas in the control group the MPR declined slightly (from 0.84 to 0.82). The difference in MPR between the two groups at 12 months also was significant. Moreover, the likelihood of an oral medication adherence rate greater than 80% (MPR 0.80 or higher) was 4.77 times greater among the intervention group than in the controls. This difference was significant despite wide confidence intervals, said Dr. Odegard.

A regression model that included demographic variables, regimen complexity, and prior oral medication adherence measures explained 70% of the change in MPR scores at 12 months. Significant predictors of a change in MPR were a low MPR during the prior period, less resistance to taking medication, and less time needed with the pharmacist.

Dr. Odegard and her associates are hoping to integrate this type of program into local pharmacy chains. Several remaining challenges include the fact that patients don’t always pick up their own medications, some might use both community and mail-order pharmacies, and some might have multiple physicians for their diabetes care.

During the question-and-answer period, Dr. Odegard remarked that such programs could provide a significant revenue stream to pharmacies. She and her colleagues are currently conducting a cost analysis.

Indeed, Safeway has been very supportive. “If a company like Safeway can take the lead on helping drive corporate policy change in helping pharmacists to do this kind of thing, if it really makes a difference, it will be huge. ... We’re really hoping that reimbursement can go to pharmacies or other health care providers who perform these services. Corporate policy change is really important to help our patients get what they need,” she said.

ORLANDO – Hip and distal fractures were found to be elevated in both women and men who had taken thiazolidinediones in an analysis of all hospitalizations for fracture in Scotland during 2000-2008.

The current product labels for both rosiglitazone and pioglitazone state that the fracture risk associated with the thiazolidinedione (TZD) class is predominantly for distal fractures occurring in women. But this study, which included an analysis using patients as their own controls to overcome treatment allocation bias, showed that the risk of hip fractures is substantially elevated in both men and women. The findings suggest that the Food and Drug Administration should consider whether a change in drug label is indicated, Dr. Helen Colhoun said at the annual scientific sessions of the American Diabetes Association.

In this analysis, no difference in risk was noted between the two individual TZDs, noted Dr. Colhoun, professor of public health at the University of Dundee, Scotland.

A clinical database identified all individuals in Scotland aged at least 40 with diagnosed type 2 diabetes, and linked those data with hospital admission records. The study comprised a total 82,027 person-years of TZD use in 35,606 users (68% rosiglitazone, 32% pioglitazone) and 6,209 fractures in 5,094 individuals.

There were 1,375 fractures among the 35,606 with diabetes who had used TZDs (3.9%) and 4,835 among the 112,100 (4.3%) who had never used TZDs. Patients who had used TZDs were significantly younger, had a higher body mass index, were less likely to have had a prior fracture, and less likely to have had recent cardiovascular disease.

After adjustment for diabetes duration, prior CVD, baseline and time-dependent exposure to other medications, and BMI, the rate of fracture per cumulative year of TZD exposure was significantly elevated for hip fracture in both men (hazard ratio 1.14) and women (HR 1.17) and for distal fracture in women (1.16).

In the self-controlled case series, the incidence rate ratio comparing the exposed and unexposed periods of time for each individual patient demonstrated a significantly increased risk of hip fractures during the period of TZD exposure for the group as a whole (incidence rate ratio 1.98) and in both men (2.23) and women (1.9). The risk for distal fractures was significantly greater for the entire group (1.4) and for women (1.57) but was not significant among men.

The incidence rate ratio by duration of TZD exposure rose from 1.45 at less than 1 year to 3.59 after 6 years of exposure, Dr. Colhoun reported.

The overall relative risk in the self-controlled case series of approximately 2 and a current prevalence of exposure of about 20% implies that about 17% of fractures in patients with diabetes could be attributable to TZDs, she noted.

Importantly, she added, this analysis assumes that the absence of TZDs would not result in more use of another agent that also increases fracture risk such as insulin (due to hypoglycemia). Studies are currently underway to assess that further and to build a predictive model of fracture in diabetes. Other studies are looking at whether reducing the TZD dose might lower the fracture risk, and whether the risk goes away when the drug is stopped. Preliminary data suggest that the risk remains high, she said.

Dr. Colhoun disclosed that she has received research support from Pfizer, Eli Lilly, Boehringer Ingelheim, AstraZeneca, and Roche. She serves on the speakers bureau for Pfizer, is a board/advisory panel member for Eli Lilly, and owns stocks/shares in Roche. The study was funded by Wellcome Trust Scottish Health Informatics Programme and the Chief Scientist’s Office, Scotland.

ORLANDO – Hip and distal fractures were found to be elevated in both women and men who had taken thiazolidinediones in an analysis of all hospitalizations for fracture in Scotland during 2000-2008.

The current product labels for both rosiglitazone and pioglitazone state that the fracture risk associated with the thiazolidinedione (TZD) class is predominantly for distal fractures occurring in women. But this study, which included an analysis using patients as their own controls to overcome treatment allocation bias, showed that the risk of hip fractures is substantially elevated in both men and women. The findings suggest that the Food and Drug Administration should consider whether a change in drug label is indicated, Dr. Helen Colhoun said at the annual scientific sessions of the American Diabetes Association.

In this analysis, no difference in risk was noted between the two individual TZDs, noted Dr. Colhoun, professor of public health at the University of Dundee, Scotland.

A clinical database identified all individuals in Scotland aged at least 40 with diagnosed type 2 diabetes, and linked those data with hospital admission records. The study comprised a total 82,027 person-years of TZD use in 35,606 users (68% rosiglitazone, 32% pioglitazone) and 6,209 fractures in 5,094 individuals.

There were 1,375 fractures among the 35,606 with diabetes who had used TZDs (3.9%) and 4,835 among the 112,100 (4.3%) who had never used TZDs. Patients who had used TZDs were significantly younger, had a higher body mass index, were less likely to have had a prior fracture, and less likely to have had recent cardiovascular disease.

After adjustment for diabetes duration, prior CVD, baseline and time-dependent exposure to other medications, and BMI, the rate of fracture per cumulative year of TZD exposure was significantly elevated for hip fracture in both men (hazard ratio 1.14) and women (HR 1.17) and for distal fracture in women (1.16).

In the self-controlled case series, the incidence rate ratio comparing the exposed and unexposed periods of time for each individual patient demonstrated a significantly increased risk of hip fractures during the period of TZD exposure for the group as a whole (incidence rate ratio 1.98) and in both men (2.23) and women (1.9). The risk for distal fractures was significantly greater for the entire group (1.4) and for women (1.57) but was not significant among men.

The incidence rate ratio by duration of TZD exposure rose from 1.45 at less than 1 year to 3.59 after 6 years of exposure, Dr. Colhoun reported.

The overall relative risk in the self-controlled case series of approximately 2 and a current prevalence of exposure of about 20% implies that about 17% of fractures in patients with diabetes could be attributable to TZDs, she noted.

Importantly, she added, this analysis assumes that the absence of TZDs would not result in more use of another agent that also increases fracture risk such as insulin (due to hypoglycemia). Studies are currently underway to assess that further and to build a predictive model of fracture in diabetes. Other studies are looking at whether reducing the TZD dose might lower the fracture risk, and whether the risk goes away when the drug is stopped. Preliminary data suggest that the risk remains high, she said.

Dr. Colhoun disclosed that she has received research support from Pfizer, Eli Lilly, Boehringer Ingelheim, AstraZeneca, and Roche. She serves on the speakers bureau for Pfizer, is a board/advisory panel member for Eli Lilly, and owns stocks/shares in Roche. The study was funded by Wellcome Trust Scottish Health Informatics Programme and the Chief Scientist’s Office, Scotland.

ORLANDO – Hip and distal fractures were found to be elevated in both women and men who had taken thiazolidinediones in an analysis of all hospitalizations for fracture in Scotland during 2000-2008.

The current product labels for both rosiglitazone and pioglitazone state that the fracture risk associated with the thiazolidinedione (TZD) class is predominantly for distal fractures occurring in women. But this study, which included an analysis using patients as their own controls to overcome treatment allocation bias, showed that the risk of hip fractures is substantially elevated in both men and women. The findings suggest that the Food and Drug Administration should consider whether a change in drug label is indicated, Dr. Helen Colhoun said at the annual scientific sessions of the American Diabetes Association.

In this analysis, no difference in risk was noted between the two individual TZDs, noted Dr. Colhoun, professor of public health at the University of Dundee, Scotland.

A clinical database identified all individuals in Scotland aged at least 40 with diagnosed type 2 diabetes, and linked those data with hospital admission records. The study comprised a total 82,027 person-years of TZD use in 35,606 users (68% rosiglitazone, 32% pioglitazone) and 6,209 fractures in 5,094 individuals.

There were 1,375 fractures among the 35,606 with diabetes who had used TZDs (3.9%) and 4,835 among the 112,100 (4.3%) who had never used TZDs. Patients who had used TZDs were significantly younger, had a higher body mass index, were less likely to have had a prior fracture, and less likely to have had recent cardiovascular disease.

After adjustment for diabetes duration, prior CVD, baseline and time-dependent exposure to other medications, and BMI, the rate of fracture per cumulative year of TZD exposure was significantly elevated for hip fracture in both men (hazard ratio 1.14) and women (HR 1.17) and for distal fracture in women (1.16).

In the self-controlled case series, the incidence rate ratio comparing the exposed and unexposed periods of time for each individual patient demonstrated a significantly increased risk of hip fractures during the period of TZD exposure for the group as a whole (incidence rate ratio 1.98) and in both men (2.23) and women (1.9). The risk for distal fractures was significantly greater for the entire group (1.4) and for women (1.57) but was not significant among men.

The incidence rate ratio by duration of TZD exposure rose from 1.45 at less than 1 year to 3.59 after 6 years of exposure, Dr. Colhoun reported.

The overall relative risk in the self-controlled case series of approximately 2 and a current prevalence of exposure of about 20% implies that about 17% of fractures in patients with diabetes could be attributable to TZDs, she noted.

Importantly, she added, this analysis assumes that the absence of TZDs would not result in more use of another agent that also increases fracture risk such as insulin (due to hypoglycemia). Studies are currently underway to assess that further and to build a predictive model of fracture in diabetes. Other studies are looking at whether reducing the TZD dose might lower the fracture risk, and whether the risk goes away when the drug is stopped. Preliminary data suggest that the risk remains high, she said.

Dr. Colhoun disclosed that she has received research support from Pfizer, Eli Lilly, Boehringer Ingelheim, AstraZeneca, and Roche. She serves on the speakers bureau for Pfizer, is a board/advisory panel member for Eli Lilly, and owns stocks/shares in Roche. The study was funded by Wellcome Trust Scottish Health Informatics Programme and the Chief Scientist’s Office, Scotland.

ORLANDO – More than 10% of non-obese adolescent girls with polycystic ovary syndrome were found to have impaired glucose tolerance in a study of 70 girls who had been referred to a specialty clinic for menstrual irregularity.

The finding suggests that all girls and women with polycystic ovary syndrome (PCOS) – not just those who are overweight or obese – should be evaluated with an oral glucose tolerance test (OGTT), said Dr. Clare A. Flannery of the Yale Multidisciplinary Adolescent PCOS Program and the department of endocrinology–internal medicine at Yale University, New Haven, Conn.

“Without an OGTT, the presence of impaired glucose metabolism is underestimated in lean adolescents with PCOS since their other parameters of insulin resistance may be in normal range. There is a need for a standardized OGTT for every adolescent with PCOS, regardless of weight,” she said.

The study was conducted at the Yale PCOS clinic, where patients referred for menstrual irregularity are seen by an endocrinologist, a gynecologist, and a nutritionist. All patients also receive a transabdominal pelvic ultrasound, androgen panel, fasting lipid testing, and a 75-g OGTT. A group of 80 patients who were enrolled in an ongoing cohort study had a mean age of 15.6 years and mean body mass index (BMI) of 31.4 kg/m2, but with a broad range of BMI from 19 to 46 kg/m2. Two-thirds of patients were white. Three-fourths of the patients had acne, nearly all had hirsutism, and more than half (57%) had acanthosis nigricans, signaling insulin resistance.

Because there are no established criteria for diagnosing PCOS in adolescence, two of three professional guidelines for diagnosing PCOS in adults were used for the study: The 1990 National Institutes of Health criteria (Blackwell Scientific Publications, 1992:377-84), which includes irregular menses and clinical or biochemical evidence of high androgens with the exclusion of other disorders, and the 2006 Androgen Excess Society (AES) criteria, which allow ultrasound findings of PCOS as a substitute for irregular menses (J. Clin. Endocrinol. Metab. 2006;91:4237-45).

(The 2003 Rotterdam criteria [Human Reproduction 2004;19:41-7] were not used because the definition is less strict and could include adolescents with hypothalamic amenorrhea, she noted.)

Ten adolescents were excluded from analysis because they were either already on metformin or had missing OGTT data. Of the remaining 70, 55 (79%) met the NIH criteria for PCOS diagnosis and 64 (91%) met the AES criteria. One patient who met both criteria was found to have impaired fasting glucose, and another who met both definitions had type 2 diabetes. Impaired glucose tolerance (IGT) was found in eight of the NIH-defined PCOS patients (14.5%) and 10 of the AES-defined group (16%).

Among the 64 who met the AES PCOS criteria, 40 were obese (BMI of 95th percentile by age or greater). Mean BMI was 36 kg/m2 for the obese group and 24 kg/m2 for the non-obese. IGT was present in 16% of both groups. Two-hour serum glucose values were higher in the obese group (117 vs. 107 mg/dL in the non-obese) but that difference, which did not meet statistical significance, was primarily because of the one outlier in the obese group who met the criteria for type 2 diabetes, Dr. Flannery noted.

“The non-obese girls were just as likely to have impaired glucose tolerance as their obese counterparts. If we had not indiscriminately applied the OGTT to all our patients, the abnormal glucose metabolism of the non-obese girls may have been missed,” she commented.

In contrast to the OGTT finding, other metabolic characteristics did appear to be driven by obesity rather than PCOS. Fasting glucose was greater – although still within normal range – among the obese patients (86 vs. 82 mg/dL), and there was a trend toward increased insulin resistance among those in the obese group, as measured by the homeostatic model assessment. Lipid abnormalities and other parameters of insulin resistance also worsened as weight increased, with both differences in C-reactive protein and high density lipid protein reaching statistical significance.

Dr. Flannery recommended that physicians use the AES guidelines for performing an OGTT in all girls and women with PCOS, regardless of age or BMI, noting that the most recent guidelines from the American Diabetes Association recommend use of OGTT only in overweight adolescents with additional risk factors. “If we had applied the ADA guidelines, we would have missed IGT in many of our adolescents,” she said.

Dr. Flannery stated that she had no financial disclosures.

ORLANDO – More than 10% of non-obese adolescent girls with polycystic ovary syndrome were found to have impaired glucose tolerance in a study of 70 girls who had been referred to a specialty clinic for menstrual irregularity.

The finding suggests that all girls and women with polycystic ovary syndrome (PCOS) – not just those who are overweight or obese – should be evaluated with an oral glucose tolerance test (OGTT), said Dr. Clare A. Flannery of the Yale Multidisciplinary Adolescent PCOS Program and the department of endocrinology–internal medicine at Yale University, New Haven, Conn.

“Without an OGTT, the presence of impaired glucose metabolism is underestimated in lean adolescents with PCOS since their other parameters of insulin resistance may be in normal range. There is a need for a standardized OGTT for every adolescent with PCOS, regardless of weight,” she said.

The study was conducted at the Yale PCOS clinic, where patients referred for menstrual irregularity are seen by an endocrinologist, a gynecologist, and a nutritionist. All patients also receive a transabdominal pelvic ultrasound, androgen panel, fasting lipid testing, and a 75-g OGTT. A group of 80 patients who were enrolled in an ongoing cohort study had a mean age of 15.6 years and mean body mass index (BMI) of 31.4 kg/m2, but with a broad range of BMI from 19 to 46 kg/m2. Two-thirds of patients were white. Three-fourths of the patients had acne, nearly all had hirsutism, and more than half (57%) had acanthosis nigricans, signaling insulin resistance.

Because there are no established criteria for diagnosing PCOS in adolescence, two of three professional guidelines for diagnosing PCOS in adults were used for the study: The 1990 National Institutes of Health criteria (Blackwell Scientific Publications, 1992:377-84), which includes irregular menses and clinical or biochemical evidence of high androgens with the exclusion of other disorders, and the 2006 Androgen Excess Society (AES) criteria, which allow ultrasound findings of PCOS as a substitute for irregular menses (J. Clin. Endocrinol. Metab. 2006;91:4237-45).

(The 2003 Rotterdam criteria [Human Reproduction 2004;19:41-7] were not used because the definition is less strict and could include adolescents with hypothalamic amenorrhea, she noted.)

Ten adolescents were excluded from analysis because they were either already on metformin or had missing OGTT data. Of the remaining 70, 55 (79%) met the NIH criteria for PCOS diagnosis and 64 (91%) met the AES criteria. One patient who met both criteria was found to have impaired fasting glucose, and another who met both definitions had type 2 diabetes. Impaired glucose tolerance (IGT) was found in eight of the NIH-defined PCOS patients (14.5%) and 10 of the AES-defined group (16%).

Among the 64 who met the AES PCOS criteria, 40 were obese (BMI of 95th percentile by age or greater). Mean BMI was 36 kg/m2 for the obese group and 24 kg/m2 for the non-obese. IGT was present in 16% of both groups. Two-hour serum glucose values were higher in the obese group (117 vs. 107 mg/dL in the non-obese) but that difference, which did not meet statistical significance, was primarily because of the one outlier in the obese group who met the criteria for type 2 diabetes, Dr. Flannery noted.

“The non-obese girls were just as likely to have impaired glucose tolerance as their obese counterparts. If we had not indiscriminately applied the OGTT to all our patients, the abnormal glucose metabolism of the non-obese girls may have been missed,” she commented.

In contrast to the OGTT finding, other metabolic characteristics did appear to be driven by obesity rather than PCOS. Fasting glucose was greater – although still within normal range – among the obese patients (86 vs. 82 mg/dL), and there was a trend toward increased insulin resistance among those in the obese group, as measured by the homeostatic model assessment. Lipid abnormalities and other parameters of insulin resistance also worsened as weight increased, with both differences in C-reactive protein and high density lipid protein reaching statistical significance.

Dr. Flannery recommended that physicians use the AES guidelines for performing an OGTT in all girls and women with PCOS, regardless of age or BMI, noting that the most recent guidelines from the American Diabetes Association recommend use of OGTT only in overweight adolescents with additional risk factors. “If we had applied the ADA guidelines, we would have missed IGT in many of our adolescents,” she said.

Dr. Flannery stated that she had no financial disclosures.

ORLANDO – More than 10% of non-obese adolescent girls with polycystic ovary syndrome were found to have impaired glucose tolerance in a study of 70 girls who had been referred to a specialty clinic for menstrual irregularity.

The finding suggests that all girls and women with polycystic ovary syndrome (PCOS) – not just those who are overweight or obese – should be evaluated with an oral glucose tolerance test (OGTT), said Dr. Clare A. Flannery of the Yale Multidisciplinary Adolescent PCOS Program and the department of endocrinology–internal medicine at Yale University, New Haven, Conn.

“Without an OGTT, the presence of impaired glucose metabolism is underestimated in lean adolescents with PCOS since their other parameters of insulin resistance may be in normal range. There is a need for a standardized OGTT for every adolescent with PCOS, regardless of weight,” she said.

The study was conducted at the Yale PCOS clinic, where patients referred for menstrual irregularity are seen by an endocrinologist, a gynecologist, and a nutritionist. All patients also receive a transabdominal pelvic ultrasound, androgen panel, fasting lipid testing, and a 75-g OGTT. A group of 80 patients who were enrolled in an ongoing cohort study had a mean age of 15.6 years and mean body mass index (BMI) of 31.4 kg/m2, but with a broad range of BMI from 19 to 46 kg/m2. Two-thirds of patients were white. Three-fourths of the patients had acne, nearly all had hirsutism, and more than half (57%) had acanthosis nigricans, signaling insulin resistance.

Because there are no established criteria for diagnosing PCOS in adolescence, two of three professional guidelines for diagnosing PCOS in adults were used for the study: The 1990 National Institutes of Health criteria (Blackwell Scientific Publications, 1992:377-84), which includes irregular menses and clinical or biochemical evidence of high androgens with the exclusion of other disorders, and the 2006 Androgen Excess Society (AES) criteria, which allow ultrasound findings of PCOS as a substitute for irregular menses (J. Clin. Endocrinol. Metab. 2006;91:4237-45).

(The 2003 Rotterdam criteria [Human Reproduction 2004;19:41-7] were not used because the definition is less strict and could include adolescents with hypothalamic amenorrhea, she noted.)

Ten adolescents were excluded from analysis because they were either already on metformin or had missing OGTT data. Of the remaining 70, 55 (79%) met the NIH criteria for PCOS diagnosis and 64 (91%) met the AES criteria. One patient who met both criteria was found to have impaired fasting glucose, and another who met both definitions had type 2 diabetes. Impaired glucose tolerance (IGT) was found in eight of the NIH-defined PCOS patients (14.5%) and 10 of the AES-defined group (16%).

Among the 64 who met the AES PCOS criteria, 40 were obese (BMI of 95th percentile by age or greater). Mean BMI was 36 kg/m2 for the obese group and 24 kg/m2 for the non-obese. IGT was present in 16% of both groups. Two-hour serum glucose values were higher in the obese group (117 vs. 107 mg/dL in the non-obese) but that difference, which did not meet statistical significance, was primarily because of the one outlier in the obese group who met the criteria for type 2 diabetes, Dr. Flannery noted.

“The non-obese girls were just as likely to have impaired glucose tolerance as their obese counterparts. If we had not indiscriminately applied the OGTT to all our patients, the abnormal glucose metabolism of the non-obese girls may have been missed,” she commented.

In contrast to the OGTT finding, other metabolic characteristics did appear to be driven by obesity rather than PCOS. Fasting glucose was greater – although still within normal range – among the obese patients (86 vs. 82 mg/dL), and there was a trend toward increased insulin resistance among those in the obese group, as measured by the homeostatic model assessment. Lipid abnormalities and other parameters of insulin resistance also worsened as weight increased, with both differences in C-reactive protein and high density lipid protein reaching statistical significance.

Dr. Flannery recommended that physicians use the AES guidelines for performing an OGTT in all girls and women with PCOS, regardless of age or BMI, noting that the most recent guidelines from the American Diabetes Association recommend use of OGTT only in overweight adolescents with additional risk factors. “If we had applied the ADA guidelines, we would have missed IGT in many of our adolescents,” she said.

Dr. Flannery stated that she had no financial disclosures.