ADA Annual Scientific Sessions 2014

SAN FRANCISCO – Among diabetes patients, the risk of cancer was significantly increased before diagnosis and 3 months after diabetes diagnosis, but not beyond that time point, results from a population-based study showed.

"Previous studies have focused on the risk of cancer following diabetes diagnosis or else compared risk of cancer between the pre- and postdiabetes period without taking into account the time period immediately following the diagnosis of diabetes," lead investigator Dr. Iliana C. Lega said in an interview prior to the annual scientific sessions of the American Diabetes Association, where the research was presented.

"We found that risk of cancer was highest in the prediabetes period and in the first 3 months following diabetes diagnosis, suggesting a role for hyperinsulinemia in the pathogenesis of cancer in the diabetes population. Furthermore, increased detection of preexisting cancers in the period following diabetes diagnosis may also contribute to the observed elevated risk of cancer," she said.

Dr. Lega, a clinician scientist with the division of endocrinology and metabolism at the University of Toronto, and her associates used population-based data from Ontario during 1997-2009 to identify incident diabetic cohorts that were matched 1:1 by age and gender to patients without diabetes. The researchers used time-varying Cox regression models to compare the relative risk of any invasive cancer between patients with and without diabetes in three separate time frames: 10 years prior to diagnosis, 3 months after diagnosis, and more than 3 months after diabetes diagnosis.

Dr. Lega and her associates evaluated cancer risk before diabetes diagnosis in 186,441 diabetes patients and matched controls and after diabetes in 51,463 diabetes patients and matched controls. Among diabetes patients, the risk of cancer was significantly increased before diabetes diagnosis (hazard ratio, 1.23) and 3 months after diabetes diagnosis (HR, 1.69), but the association abated beyond 3 months after diabetes diagnosis.

The findings "support a role for hyperinsulinemia in the relationship between diabetes and cancer, over surveillance bias," the researchers concluded in their abstract. "While further research is needed, these results support enhanced cancer screening in patients with diagnoses of prediabetes or diabetes."

Dr. Lega acknowledged certain limitations of the study, including the fact that other risk factors for cancer such as clinical features, smoking history, chronic alcoholism, and family history could not be taken into account.

Dr. Lega said that she had no relevant financial conflicts to disclose.

dbrunk@frontlinemedcom.com

On Twitter @dougbrunk 

SAN FRANCISCO – Among diabetes patients, the risk of cancer was significantly increased before diagnosis and 3 months after diabetes diagnosis, but not beyond that time point, results from a population-based study showed.

"Previous studies have focused on the risk of cancer following diabetes diagnosis or else compared risk of cancer between the pre- and postdiabetes period without taking into account the time period immediately following the diagnosis of diabetes," lead investigator Dr. Iliana C. Lega said in an interview prior to the annual scientific sessions of the American Diabetes Association, where the research was presented.

"We found that risk of cancer was highest in the prediabetes period and in the first 3 months following diabetes diagnosis, suggesting a role for hyperinsulinemia in the pathogenesis of cancer in the diabetes population. Furthermore, increased detection of preexisting cancers in the period following diabetes diagnosis may also contribute to the observed elevated risk of cancer," she said.

Dr. Lega, a clinician scientist with the division of endocrinology and metabolism at the University of Toronto, and her associates used population-based data from Ontario during 1997-2009 to identify incident diabetic cohorts that were matched 1:1 by age and gender to patients without diabetes. The researchers used time-varying Cox regression models to compare the relative risk of any invasive cancer between patients with and without diabetes in three separate time frames: 10 years prior to diagnosis, 3 months after diagnosis, and more than 3 months after diabetes diagnosis.

Dr. Lega and her associates evaluated cancer risk before diabetes diagnosis in 186,441 diabetes patients and matched controls and after diabetes in 51,463 diabetes patients and matched controls. Among diabetes patients, the risk of cancer was significantly increased before diabetes diagnosis (hazard ratio, 1.23) and 3 months after diabetes diagnosis (HR, 1.69), but the association abated beyond 3 months after diabetes diagnosis.

The findings "support a role for hyperinsulinemia in the relationship between diabetes and cancer, over surveillance bias," the researchers concluded in their abstract. "While further research is needed, these results support enhanced cancer screening in patients with diagnoses of prediabetes or diabetes."

Dr. Lega acknowledged certain limitations of the study, including the fact that other risk factors for cancer such as clinical features, smoking history, chronic alcoholism, and family history could not be taken into account.

Dr. Lega said that she had no relevant financial conflicts to disclose.

dbrunk@frontlinemedcom.com

On Twitter @dougbrunk 

SAN FRANCISCO – Among diabetes patients, the risk of cancer was significantly increased before diagnosis and 3 months after diabetes diagnosis, but not beyond that time point, results from a population-based study showed.

"Previous studies have focused on the risk of cancer following diabetes diagnosis or else compared risk of cancer between the pre- and postdiabetes period without taking into account the time period immediately following the diagnosis of diabetes," lead investigator Dr. Iliana C. Lega said in an interview prior to the annual scientific sessions of the American Diabetes Association, where the research was presented.

"We found that risk of cancer was highest in the prediabetes period and in the first 3 months following diabetes diagnosis, suggesting a role for hyperinsulinemia in the pathogenesis of cancer in the diabetes population. Furthermore, increased detection of preexisting cancers in the period following diabetes diagnosis may also contribute to the observed elevated risk of cancer," she said.

Dr. Lega, a clinician scientist with the division of endocrinology and metabolism at the University of Toronto, and her associates used population-based data from Ontario during 1997-2009 to identify incident diabetic cohorts that were matched 1:1 by age and gender to patients without diabetes. The researchers used time-varying Cox regression models to compare the relative risk of any invasive cancer between patients with and without diabetes in three separate time frames: 10 years prior to diagnosis, 3 months after diagnosis, and more than 3 months after diabetes diagnosis.

Dr. Lega and her associates evaluated cancer risk before diabetes diagnosis in 186,441 diabetes patients and matched controls and after diabetes in 51,463 diabetes patients and matched controls. Among diabetes patients, the risk of cancer was significantly increased before diabetes diagnosis (hazard ratio, 1.23) and 3 months after diabetes diagnosis (HR, 1.69), but the association abated beyond 3 months after diabetes diagnosis.

The findings "support a role for hyperinsulinemia in the relationship between diabetes and cancer, over surveillance bias," the researchers concluded in their abstract. "While further research is needed, these results support enhanced cancer screening in patients with diagnoses of prediabetes or diabetes."

Dr. Lega acknowledged certain limitations of the study, including the fact that other risk factors for cancer such as clinical features, smoking history, chronic alcoholism, and family history could not be taken into account.

Dr. Lega said that she had no relevant financial conflicts to disclose.

dbrunk@frontlinemedcom.com

On Twitter @dougbrunk 

SAN FRANCISCO – The U.S. incidence of adult diabetes doubled between 1980 and 2008 but has fallen a bit since then, which may be a sign that the diabetes epidemic is abating, according to Linda S. Geiss a health statistician with the Centers for Disease Control and Prevention.

This potentially good news does not apply to everyone, however. Both the incidence and prevalence of diabetes continue to increase in Hispanics, non-Hispanic blacks, and adults with less than a high school education, Ms. Geiss and her associates reported at the annual scientific session of the American Diabetes Association.

They studied annual data from the 1980-2012 National Health Interview Surveys to identify diabetes trends in U.S. residents aged 20-79 years. The age-adjusted prevalence and incidence of diagnosed diabetes changed little in the 1980s but each doubled between 1990 and 2008, she said.

During that time period, the U.S. population became older, less white, and better educated, she noted.

The age-adjusted incidence of diabetes increased from approximately 4/1,000 people in 1980 to more than 9/1,000 in 2008 and then declined to less than 8/1,000 in 2012, a statistically insignificant decrease after 2008. The age-adjusted prevalence of diabetes increased from approximately 4/100 people in 1980 to approximately 8/100 in 2008, with a slight increase after that, "although at a slower rate of pace," Ms. Geiss said.

Even when incidence declines, prevalence can continue to rise if the number of deaths among people with diabetes is smaller than the number of new diagnoses.

"Overall and for some subpopulations, incidence and/or prevalence slowed or plateaued around 2008. Some groups slowed earlier in the 2000s," she said. "After a steady 15- to 20-year increase in prevalence and incidence, we are seeing the first signs that the growth may be slowing or abating. However, given the uncertainty about the reasons behind these changes, future trends are uncertain. Given the large burden of diabetes in the United States, we need to sustain efforts to prevent diabetes and its complications," especially among population groups whose incidence and prevalence rates continue to climb.

Throughout the time period studied, the incidence and prevalence of diabetes were lowest among people aged 20-44 years and highest among people aged 65-79 years. For those two age groups, the incidence of diabetes increased throughout the time period. For middle-aged people of 45-64 years, the incidence plateaued in 2002. The prevalence of diabetes plateaued in 2008 for the middle-aged group and in 2003 for the oldest age group.

Incidence rates for men started to exceed those for women around 1997, continued increasing until 2008, and then declined, though not significantly. The incidence for women increased throughout the time period studied. The prevalence of diabetes plateaued for men in 2001 and for women in 2008.

For adults with a high school education, the incidence of diabetes increased until 2008 and then decreased insignificantly. For other educational levels (more than or less than a high school education), the incidence increased throughout the period. The prevalence slowed among people with more than a high school education in 2000, but increased throughout the time period for the other educational levels.

The incidence and prevalence of diabetes were higher among Hispanics and non-Hispanic blacks than among whites between 1997 and 2012. For whites, the incidence increased from approximately 5/1,000 people in 1997 to 8/1,000 in 2008, then decreased insignificantly to 6/1,000 in 2012. The prevalence of diabetes in whites slowed its rate of increase starting in 2005, Ms. Geiss reported.

"With these cross-sectional data, you can’t determine the reasons behind trend changes," she said. The nationally representative data spanning 3 decades give strength to the findings, but the surveys did not include institutionalized residents or people with undiagnosed diabetes. The study could not distinguish trends for type 1 vs. type 2 diabetes.

One physician in the audience asked if the global financial crisis in 2008 may have been a factor in slowing the diabetes epidemic.

Ms. Geiss said she hadn’t considered that possible explanation. Other factors that may have affected the incidence of diabetes include the adoption of hemoglobin A_1c (HbA_1c) for the diagnosis of diabetes, she speculated. "We know that HbA_1c tends to identify fewer people who have hyperglycemia," she said. Also, U.S. obesity rates have not increased since 2003-2004, and a couple of separate studies have reported declining caloric intake by the U.S. population. Each of these factors may be "prominent drivers" of the slowing incidence of diabetes, she said.

Ms. Geiss reported having no financial disclosures.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – The U.S. incidence of adult diabetes doubled between 1980 and 2008 but has fallen a bit since then, which may be a sign that the diabetes epidemic is abating, according to Linda S. Geiss a health statistician with the Centers for Disease Control and Prevention.

This potentially good news does not apply to everyone, however. Both the incidence and prevalence of diabetes continue to increase in Hispanics, non-Hispanic blacks, and adults with less than a high school education, Ms. Geiss and her associates reported at the annual scientific session of the American Diabetes Association.

They studied annual data from the 1980-2012 National Health Interview Surveys to identify diabetes trends in U.S. residents aged 20-79 years. The age-adjusted prevalence and incidence of diagnosed diabetes changed little in the 1980s but each doubled between 1990 and 2008, she said.

During that time period, the U.S. population became older, less white, and better educated, she noted.

The age-adjusted incidence of diabetes increased from approximately 4/1,000 people in 1980 to more than 9/1,000 in 2008 and then declined to less than 8/1,000 in 2012, a statistically insignificant decrease after 2008. The age-adjusted prevalence of diabetes increased from approximately 4/100 people in 1980 to approximately 8/100 in 2008, with a slight increase after that, "although at a slower rate of pace," Ms. Geiss said.

Even when incidence declines, prevalence can continue to rise if the number of deaths among people with diabetes is smaller than the number of new diagnoses.

"Overall and for some subpopulations, incidence and/or prevalence slowed or plateaued around 2008. Some groups slowed earlier in the 2000s," she said. "After a steady 15- to 20-year increase in prevalence and incidence, we are seeing the first signs that the growth may be slowing or abating. However, given the uncertainty about the reasons behind these changes, future trends are uncertain. Given the large burden of diabetes in the United States, we need to sustain efforts to prevent diabetes and its complications," especially among population groups whose incidence and prevalence rates continue to climb.

Throughout the time period studied, the incidence and prevalence of diabetes were lowest among people aged 20-44 years and highest among people aged 65-79 years. For those two age groups, the incidence of diabetes increased throughout the time period. For middle-aged people of 45-64 years, the incidence plateaued in 2002. The prevalence of diabetes plateaued in 2008 for the middle-aged group and in 2003 for the oldest age group.

Incidence rates for men started to exceed those for women around 1997, continued increasing until 2008, and then declined, though not significantly. The incidence for women increased throughout the time period studied. The prevalence of diabetes plateaued for men in 2001 and for women in 2008.

For adults with a high school education, the incidence of diabetes increased until 2008 and then decreased insignificantly. For other educational levels (more than or less than a high school education), the incidence increased throughout the period. The prevalence slowed among people with more than a high school education in 2000, but increased throughout the time period for the other educational levels.

The incidence and prevalence of diabetes were higher among Hispanics and non-Hispanic blacks than among whites between 1997 and 2012. For whites, the incidence increased from approximately 5/1,000 people in 1997 to 8/1,000 in 2008, then decreased insignificantly to 6/1,000 in 2012. The prevalence of diabetes in whites slowed its rate of increase starting in 2005, Ms. Geiss reported.

"With these cross-sectional data, you can’t determine the reasons behind trend changes," she said. The nationally representative data spanning 3 decades give strength to the findings, but the surveys did not include institutionalized residents or people with undiagnosed diabetes. The study could not distinguish trends for type 1 vs. type 2 diabetes.

One physician in the audience asked if the global financial crisis in 2008 may have been a factor in slowing the diabetes epidemic.

Ms. Geiss said she hadn’t considered that possible explanation. Other factors that may have affected the incidence of diabetes include the adoption of hemoglobin A_1c (HbA_1c) for the diagnosis of diabetes, she speculated. "We know that HbA_1c tends to identify fewer people who have hyperglycemia," she said. Also, U.S. obesity rates have not increased since 2003-2004, and a couple of separate studies have reported declining caloric intake by the U.S. population. Each of these factors may be "prominent drivers" of the slowing incidence of diabetes, she said.

Ms. Geiss reported having no financial disclosures.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – The U.S. incidence of adult diabetes doubled between 1980 and 2008 but has fallen a bit since then, which may be a sign that the diabetes epidemic is abating, according to Linda S. Geiss a health statistician with the Centers for Disease Control and Prevention.

This potentially good news does not apply to everyone, however. Both the incidence and prevalence of diabetes continue to increase in Hispanics, non-Hispanic blacks, and adults with less than a high school education, Ms. Geiss and her associates reported at the annual scientific session of the American Diabetes Association.

They studied annual data from the 1980-2012 National Health Interview Surveys to identify diabetes trends in U.S. residents aged 20-79 years. The age-adjusted prevalence and incidence of diagnosed diabetes changed little in the 1980s but each doubled between 1990 and 2008, she said.

During that time period, the U.S. population became older, less white, and better educated, she noted.

The age-adjusted incidence of diabetes increased from approximately 4/1,000 people in 1980 to more than 9/1,000 in 2008 and then declined to less than 8/1,000 in 2012, a statistically insignificant decrease after 2008. The age-adjusted prevalence of diabetes increased from approximately 4/100 people in 1980 to approximately 8/100 in 2008, with a slight increase after that, "although at a slower rate of pace," Ms. Geiss said.

Even when incidence declines, prevalence can continue to rise if the number of deaths among people with diabetes is smaller than the number of new diagnoses.

"Overall and for some subpopulations, incidence and/or prevalence slowed or plateaued around 2008. Some groups slowed earlier in the 2000s," she said. "After a steady 15- to 20-year increase in prevalence and incidence, we are seeing the first signs that the growth may be slowing or abating. However, given the uncertainty about the reasons behind these changes, future trends are uncertain. Given the large burden of diabetes in the United States, we need to sustain efforts to prevent diabetes and its complications," especially among population groups whose incidence and prevalence rates continue to climb.

Throughout the time period studied, the incidence and prevalence of diabetes were lowest among people aged 20-44 years and highest among people aged 65-79 years. For those two age groups, the incidence of diabetes increased throughout the time period. For middle-aged people of 45-64 years, the incidence plateaued in 2002. The prevalence of diabetes plateaued in 2008 for the middle-aged group and in 2003 for the oldest age group.

Incidence rates for men started to exceed those for women around 1997, continued increasing until 2008, and then declined, though not significantly. The incidence for women increased throughout the time period studied. The prevalence of diabetes plateaued for men in 2001 and for women in 2008.

For adults with a high school education, the incidence of diabetes increased until 2008 and then decreased insignificantly. For other educational levels (more than or less than a high school education), the incidence increased throughout the period. The prevalence slowed among people with more than a high school education in 2000, but increased throughout the time period for the other educational levels.

The incidence and prevalence of diabetes were higher among Hispanics and non-Hispanic blacks than among whites between 1997 and 2012. For whites, the incidence increased from approximately 5/1,000 people in 1997 to 8/1,000 in 2008, then decreased insignificantly to 6/1,000 in 2012. The prevalence of diabetes in whites slowed its rate of increase starting in 2005, Ms. Geiss reported.

"With these cross-sectional data, you can’t determine the reasons behind trend changes," she said. The nationally representative data spanning 3 decades give strength to the findings, but the surveys did not include institutionalized residents or people with undiagnosed diabetes. The study could not distinguish trends for type 1 vs. type 2 diabetes.

One physician in the audience asked if the global financial crisis in 2008 may have been a factor in slowing the diabetes epidemic.

Ms. Geiss said she hadn’t considered that possible explanation. Other factors that may have affected the incidence of diabetes include the adoption of hemoglobin A_1c (HbA_1c) for the diagnosis of diabetes, she speculated. "We know that HbA_1c tends to identify fewer people who have hyperglycemia," she said. Also, U.S. obesity rates have not increased since 2003-2004, and a couple of separate studies have reported declining caloric intake by the U.S. population. Each of these factors may be "prominent drivers" of the slowing incidence of diabetes, she said.

Ms. Geiss reported having no financial disclosures.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Imeglimin, the first in a new tetrahydrotriazine-containing class of oral glucose–lowering agents, markedly increased glucose-dependent insulin sensitivity and improved beta-cell function, a small randomized trial showed.

The agent, developed by Lyon, France–based Poxel, is currently in phase IIb development in the United States and Europe, regulates mitochondrial bioenergetics, and targets the liver, muscle, and pancreas, Dr. Michael Roden said at the annual scientific sessions of the American Diabetes Association. Imeglimin "has been shown to improve liver and muscle insulin sensitivity, and it is also efficient in untreated, metformin- or sitagliptin-treated type 2 diabetes without safety concerns," said Dr. Roden, director of the German Diabetes Center, Düsseldorf.

The aim of the current study was to examine the subacute effect of imeglimin on glucose-stimulated insulin secretion (GSIS) in vivo during hyperglycemic glucose clamp tests. The researchers enrolled 30 patients with type 2 diabetes who were either drug naive or on previous metformin monotherapy for at least 12 weeks. Their mean age was 61 years, their mean body mass index was 29.7 kg/m², and "they were reasonably well controlled at baseline, with a hemoglobin A_1c of 6.5%-7.5% at the first screening visit," Dr. Roden said.

After a 2-week metformin wash-out period, he and his associates compared imeglimin 1,500 mg twice daily for 1 week with placebo. A 90-minute hyperglycemic clamp was performed with administration of arginine at 45 minutes to assess the maximal response. The primary endpoint of the study was insulin secretion assessed by total insulin response given as the incremental area under the curve from 0 to 45 minutes as well as insulin secretion rate (ISR) calculated from C-peptide deconvolution. First-phase insulin sensitivity was defined as that which occurred in the first 8 minutes of administration while second-phase insulin sensitivity was defined as that which occurred after 25-45 minutes of administration.

Beta-cell glucose sensitivity was measured as the ratio of incremental ISR at steady state to mean incremental glucose levels during the same period, hepatic insulin extraction was measured by mathematical modeling from insulin and C-peptide concentrations, and insulin clearance was measured as a ratio of second-phase ISR and the corresponding area under the curve of insulin concentrations.

Dr. Roden reported that in drug-naive or metformin-treated patients with type 2 diabetes, imeglimin treatment for 7 days raised insulin response to glucose by 112% (P = .03), increased first-phase ISR by 110% (P = .034), increased second-phase ISR by 30% (P = .031), and improved beta-cell glucose sensitivity by 36% (P = .034). It also tended to decrease hepatic insulin extraction. "Imeglimin improves beta-cell function in humans, which may explain its glucose-lowering effect in longer-lasting trials," he said.

Poxel supported the study. Two of the study authors are employed by the company. Dr. Roden disclosed that he serves on advisory boards for Poxel.

dbrunk@frontlinemedcom.com

SAN FRANCISCO – Imeglimin, the first in a new tetrahydrotriazine-containing class of oral glucose–lowering agents, markedly increased glucose-dependent insulin sensitivity and improved beta-cell function, a small randomized trial showed.

The agent, developed by Lyon, France–based Poxel, is currently in phase IIb development in the United States and Europe, regulates mitochondrial bioenergetics, and targets the liver, muscle, and pancreas, Dr. Michael Roden said at the annual scientific sessions of the American Diabetes Association. Imeglimin "has been shown to improve liver and muscle insulin sensitivity, and it is also efficient in untreated, metformin- or sitagliptin-treated type 2 diabetes without safety concerns," said Dr. Roden, director of the German Diabetes Center, Düsseldorf.

The aim of the current study was to examine the subacute effect of imeglimin on glucose-stimulated insulin secretion (GSIS) in vivo during hyperglycemic glucose clamp tests. The researchers enrolled 30 patients with type 2 diabetes who were either drug naive or on previous metformin monotherapy for at least 12 weeks. Their mean age was 61 years, their mean body mass index was 29.7 kg/m², and "they were reasonably well controlled at baseline, with a hemoglobin A_1c of 6.5%-7.5% at the first screening visit," Dr. Roden said.

After a 2-week metformin wash-out period, he and his associates compared imeglimin 1,500 mg twice daily for 1 week with placebo. A 90-minute hyperglycemic clamp was performed with administration of arginine at 45 minutes to assess the maximal response. The primary endpoint of the study was insulin secretion assessed by total insulin response given as the incremental area under the curve from 0 to 45 minutes as well as insulin secretion rate (ISR) calculated from C-peptide deconvolution. First-phase insulin sensitivity was defined as that which occurred in the first 8 minutes of administration while second-phase insulin sensitivity was defined as that which occurred after 25-45 minutes of administration.

Beta-cell glucose sensitivity was measured as the ratio of incremental ISR at steady state to mean incremental glucose levels during the same period, hepatic insulin extraction was measured by mathematical modeling from insulin and C-peptide concentrations, and insulin clearance was measured as a ratio of second-phase ISR and the corresponding area under the curve of insulin concentrations.

Dr. Roden reported that in drug-naive or metformin-treated patients with type 2 diabetes, imeglimin treatment for 7 days raised insulin response to glucose by 112% (P = .03), increased first-phase ISR by 110% (P = .034), increased second-phase ISR by 30% (P = .031), and improved beta-cell glucose sensitivity by 36% (P = .034). It also tended to decrease hepatic insulin extraction. "Imeglimin improves beta-cell function in humans, which may explain its glucose-lowering effect in longer-lasting trials," he said.

Poxel supported the study. Two of the study authors are employed by the company. Dr. Roden disclosed that he serves on advisory boards for Poxel.

dbrunk@frontlinemedcom.com

SAN FRANCISCO – Imeglimin, the first in a new tetrahydrotriazine-containing class of oral glucose–lowering agents, markedly increased glucose-dependent insulin sensitivity and improved beta-cell function, a small randomized trial showed.

The agent, developed by Lyon, France–based Poxel, is currently in phase IIb development in the United States and Europe, regulates mitochondrial bioenergetics, and targets the liver, muscle, and pancreas, Dr. Michael Roden said at the annual scientific sessions of the American Diabetes Association. Imeglimin "has been shown to improve liver and muscle insulin sensitivity, and it is also efficient in untreated, metformin- or sitagliptin-treated type 2 diabetes without safety concerns," said Dr. Roden, director of the German Diabetes Center, Düsseldorf.

The aim of the current study was to examine the subacute effect of imeglimin on glucose-stimulated insulin secretion (GSIS) in vivo during hyperglycemic glucose clamp tests. The researchers enrolled 30 patients with type 2 diabetes who were either drug naive or on previous metformin monotherapy for at least 12 weeks. Their mean age was 61 years, their mean body mass index was 29.7 kg/m², and "they were reasonably well controlled at baseline, with a hemoglobin A_1c of 6.5%-7.5% at the first screening visit," Dr. Roden said.

After a 2-week metformin wash-out period, he and his associates compared imeglimin 1,500 mg twice daily for 1 week with placebo. A 90-minute hyperglycemic clamp was performed with administration of arginine at 45 minutes to assess the maximal response. The primary endpoint of the study was insulin secretion assessed by total insulin response given as the incremental area under the curve from 0 to 45 minutes as well as insulin secretion rate (ISR) calculated from C-peptide deconvolution. First-phase insulin sensitivity was defined as that which occurred in the first 8 minutes of administration while second-phase insulin sensitivity was defined as that which occurred after 25-45 minutes of administration.

Beta-cell glucose sensitivity was measured as the ratio of incremental ISR at steady state to mean incremental glucose levels during the same period, hepatic insulin extraction was measured by mathematical modeling from insulin and C-peptide concentrations, and insulin clearance was measured as a ratio of second-phase ISR and the corresponding area under the curve of insulin concentrations.

Dr. Roden reported that in drug-naive or metformin-treated patients with type 2 diabetes, imeglimin treatment for 7 days raised insulin response to glucose by 112% (P = .03), increased first-phase ISR by 110% (P = .034), increased second-phase ISR by 30% (P = .031), and improved beta-cell glucose sensitivity by 36% (P = .034). It also tended to decrease hepatic insulin extraction. "Imeglimin improves beta-cell function in humans, which may explain its glucose-lowering effect in longer-lasting trials," he said.

Poxel supported the study. Two of the study authors are employed by the company. Dr. Roden disclosed that he serves on advisory boards for Poxel.

dbrunk@frontlinemedcom.com

SAN FRANCISCO – Patients with type 2 diabetes who were insulin deficient were 36 times more likely to develop severe hypoglycemia on treatment and to fail to get hemoglobin A_1c levels below 6% after starting intensive therapy in a retrospective analysis of data on 1,401 patients.

In addition, patients with any of four antibodies against islet cells at baseline were 4-12 times more likely to develop severe hypoglycemia and fail to reach the glycemic target on intensive therapy, compared with patients without those biomarkers, Dr. Lisa Chow reported at the annual scientific sessions of the American Diabetes Association.

The study defined insulin deficiency as a fasting C-peptide level no higher than 0.45 ng/mL. Investigators measured levels of three of the autoantibodies and considered the fourth – insulin autoantibody (IAA) – to be present if the patient used insulin at baseline.

In unadjusted analyses, the likelihood of severe hypoglycemia and failure to reach the glycemic target was four times higher in patients with IAA or antibodies against glutamic acid decarboxylase (GAD) or zinc transporter 8 (ZnT8), compared with patients without those autoantibodies. Severe hypoglycemia and failure to get hemoglobin A_1c (HbA_1c) below 6% were 11 times more likely in patients with antibodies against tyrosine phosphatase–related islet antigen 2 (IA-2), reported Dr. Chow, an endocrinologist at the University of Minnesota, Minneapolis.

She and her associates studied data on 1,401 patients who were randomized to the intensive glycemic control arm of the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial. They compared 326 patients ("cases") who developed severe hypoglycemia and whose HbA_1c levels did not reach the target of less than 6% with 1,075 control patients (matched 4:1 by age, race, and body mass index) who did not have severe hypoglycemia and who reached the HbA_1c target.

Overall, 17% of cases and 1% of controls were insulin deficient. Excluding IAA (present in 62% of cases and 26% of controls), the most common islet antibody was GAD, in 17% of cases and 6% of controls. IA-2 was present in 4% of cases and 0.4% of controls, and ZnT8 was present in 2% of cases and 0.7% of controls.

"We propose that measuring C-peptide and GAD antibodies in patients with type 2 diabetes may serve as a biomarker to help tailor and individualize therapy" by predicting which patients are more likely to have adverse outcomes on intensive glycemic treatment, Dr. Chow said.

At baseline, the case group was significantly more likely to use insulin, compared with controls (89% and 24%, respectively) and less likely to be male (50% vs. 59%). The case patients had a higher HbA_1c at baseline, compared with controls (8.54% vs. 8.08%, respectively) and a longer duration of diabetes (15 vs. 9 years).

After adjusting for the effects of age and body mass index, the combined outcome of severe hypoglycemia and failure to reach the glycemic target was 32 times more likely in patients with insulin deficiency at baseline, 3 times more likely in patients with GAD or ZnT8, 5 times more likely in patients with IAA, and 9 times more likely in patients with IA-2, compared with patients without those markers, Dr. Chow said.

A third analysis removed data for any case patients who died and their associated controls and found that severe hypoglycemia and failure to reach the glycemic target was 30 times more likely in patients with insulin deficiency at baseline, 4 times more likely in patients with GAD or ZnT8, 5 times more likely in patients with IAA, and 12 times more likely in patients with IA-2, compared with patients without those markers.

The study defined severe hypoglycemia as any hypoglycemia requiring assistance. Patients who developed severe hypoglycemia were four times less likely to get their HbA_1c levels below 6%.

Dr. Chow reported having no financial disclosures. The National Institutes of Health funded the study.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Patients with type 2 diabetes who were insulin deficient were 36 times more likely to develop severe hypoglycemia on treatment and to fail to get hemoglobin A_1c levels below 6% after starting intensive therapy in a retrospective analysis of data on 1,401 patients.

In addition, patients with any of four antibodies against islet cells at baseline were 4-12 times more likely to develop severe hypoglycemia and fail to reach the glycemic target on intensive therapy, compared with patients without those biomarkers, Dr. Lisa Chow reported at the annual scientific sessions of the American Diabetes Association.

The study defined insulin deficiency as a fasting C-peptide level no higher than 0.45 ng/mL. Investigators measured levels of three of the autoantibodies and considered the fourth – insulin autoantibody (IAA) – to be present if the patient used insulin at baseline.

In unadjusted analyses, the likelihood of severe hypoglycemia and failure to reach the glycemic target was four times higher in patients with IAA or antibodies against glutamic acid decarboxylase (GAD) or zinc transporter 8 (ZnT8), compared with patients without those autoantibodies. Severe hypoglycemia and failure to get hemoglobin A_1c (HbA_1c) below 6% were 11 times more likely in patients with antibodies against tyrosine phosphatase–related islet antigen 2 (IA-2), reported Dr. Chow, an endocrinologist at the University of Minnesota, Minneapolis.

She and her associates studied data on 1,401 patients who were randomized to the intensive glycemic control arm of the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial. They compared 326 patients ("cases") who developed severe hypoglycemia and whose HbA_1c levels did not reach the target of less than 6% with 1,075 control patients (matched 4:1 by age, race, and body mass index) who did not have severe hypoglycemia and who reached the HbA_1c target.

Overall, 17% of cases and 1% of controls were insulin deficient. Excluding IAA (present in 62% of cases and 26% of controls), the most common islet antibody was GAD, in 17% of cases and 6% of controls. IA-2 was present in 4% of cases and 0.4% of controls, and ZnT8 was present in 2% of cases and 0.7% of controls.

"We propose that measuring C-peptide and GAD antibodies in patients with type 2 diabetes may serve as a biomarker to help tailor and individualize therapy" by predicting which patients are more likely to have adverse outcomes on intensive glycemic treatment, Dr. Chow said.

At baseline, the case group was significantly more likely to use insulin, compared with controls (89% and 24%, respectively) and less likely to be male (50% vs. 59%). The case patients had a higher HbA_1c at baseline, compared with controls (8.54% vs. 8.08%, respectively) and a longer duration of diabetes (15 vs. 9 years).

After adjusting for the effects of age and body mass index, the combined outcome of severe hypoglycemia and failure to reach the glycemic target was 32 times more likely in patients with insulin deficiency at baseline, 3 times more likely in patients with GAD or ZnT8, 5 times more likely in patients with IAA, and 9 times more likely in patients with IA-2, compared with patients without those markers, Dr. Chow said.

A third analysis removed data for any case patients who died and their associated controls and found that severe hypoglycemia and failure to reach the glycemic target was 30 times more likely in patients with insulin deficiency at baseline, 4 times more likely in patients with GAD or ZnT8, 5 times more likely in patients with IAA, and 12 times more likely in patients with IA-2, compared with patients without those markers.

The study defined severe hypoglycemia as any hypoglycemia requiring assistance. Patients who developed severe hypoglycemia were four times less likely to get their HbA_1c levels below 6%.

Dr. Chow reported having no financial disclosures. The National Institutes of Health funded the study.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Patients with type 2 diabetes who were insulin deficient were 36 times more likely to develop severe hypoglycemia on treatment and to fail to get hemoglobin A_1c levels below 6% after starting intensive therapy in a retrospective analysis of data on 1,401 patients.

In addition, patients with any of four antibodies against islet cells at baseline were 4-12 times more likely to develop severe hypoglycemia and fail to reach the glycemic target on intensive therapy, compared with patients without those biomarkers, Dr. Lisa Chow reported at the annual scientific sessions of the American Diabetes Association.

The study defined insulin deficiency as a fasting C-peptide level no higher than 0.45 ng/mL. Investigators measured levels of three of the autoantibodies and considered the fourth – insulin autoantibody (IAA) – to be present if the patient used insulin at baseline.

In unadjusted analyses, the likelihood of severe hypoglycemia and failure to reach the glycemic target was four times higher in patients with IAA or antibodies against glutamic acid decarboxylase (GAD) or zinc transporter 8 (ZnT8), compared with patients without those autoantibodies. Severe hypoglycemia and failure to get hemoglobin A_1c (HbA_1c) below 6% were 11 times more likely in patients with antibodies against tyrosine phosphatase–related islet antigen 2 (IA-2), reported Dr. Chow, an endocrinologist at the University of Minnesota, Minneapolis.

She and her associates studied data on 1,401 patients who were randomized to the intensive glycemic control arm of the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial. They compared 326 patients ("cases") who developed severe hypoglycemia and whose HbA_1c levels did not reach the target of less than 6% with 1,075 control patients (matched 4:1 by age, race, and body mass index) who did not have severe hypoglycemia and who reached the HbA_1c target.

Overall, 17% of cases and 1% of controls were insulin deficient. Excluding IAA (present in 62% of cases and 26% of controls), the most common islet antibody was GAD, in 17% of cases and 6% of controls. IA-2 was present in 4% of cases and 0.4% of controls, and ZnT8 was present in 2% of cases and 0.7% of controls.

"We propose that measuring C-peptide and GAD antibodies in patients with type 2 diabetes may serve as a biomarker to help tailor and individualize therapy" by predicting which patients are more likely to have adverse outcomes on intensive glycemic treatment, Dr. Chow said.

At baseline, the case group was significantly more likely to use insulin, compared with controls (89% and 24%, respectively) and less likely to be male (50% vs. 59%). The case patients had a higher HbA_1c at baseline, compared with controls (8.54% vs. 8.08%, respectively) and a longer duration of diabetes (15 vs. 9 years).

After adjusting for the effects of age and body mass index, the combined outcome of severe hypoglycemia and failure to reach the glycemic target was 32 times more likely in patients with insulin deficiency at baseline, 3 times more likely in patients with GAD or ZnT8, 5 times more likely in patients with IAA, and 9 times more likely in patients with IA-2, compared with patients without those markers, Dr. Chow said.

A third analysis removed data for any case patients who died and their associated controls and found that severe hypoglycemia and failure to reach the glycemic target was 30 times more likely in patients with insulin deficiency at baseline, 4 times more likely in patients with GAD or ZnT8, 5 times more likely in patients with IAA, and 12 times more likely in patients with IA-2, compared with patients without those markers.

The study defined severe hypoglycemia as any hypoglycemia requiring assistance. Patients who developed severe hypoglycemia were four times less likely to get their HbA_1c levels below 6%.

Dr. Chow reported having no financial disclosures. The National Institutes of Health funded the study.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Getting patients to abandon unhealthful behaviors to better control diabetes can be hard, and sometimes the physician is the problem, according to behavioral psychologist William H. Polonsky, Ph.D.

"I think we’re a little too in love with the idea of behavior change," Dr. Polonsky said in the Richard R. Rubin Award lecture at the annual scientific sessions of the American Diabetes Association. "The major thing I’ve learned from my patients is we need to look in the mirror" to figure out why efforts at behavior change aren’t working.

Certainly, healthy behaviors are incredibly helpful. Patients who reported a high level of disease self-management were nearly 40% less likely to die over a 12-year period in a study of 340 patients with type 2 diabetes (Diabetes Care 2014; 37:1604-12). But when behavior change isn’t happening, think about three common mistakes that clinicians make in the process, said Dr. Polonsky, president of the nonprofit Behavioral Diabetes Institute, San Diego:

1. Not taking enough time. Don’t rush patients to change their unhealthful behavior too soon, Dr. Polonsky advised. Make certain your patient believes that the selected behavior change is worthwhile.

"I push my patients way too hard, and I push them when I shouldn’t be pushing them," he said. "It’s this urge we have because we care and probably because we’re overtrained as problem-solvers."

Many clinicians now employ motivational interviewing, for example, which is supposed to end with formulation of concrete, realistic goals for change. However, in reviewing four fairly large randomized, controlled trials published in 2010-2013 that compared motivational interviewing techniques to usual care in managing patients with diabetes, Dr. Polonsky found that the change in hemoglobin A_1c (HbA_1c) levels differed between groups by an average of only 0.1%. Two other studies found that using motivational interviewing was significantly less effective than diabetes education in changing HbA_1c levels.

Perhaps, the emphasis on formulating a concrete goal does not allow enough time to convince a patient that a specific behavior change would be worthwhile, he said.

2. Not focusing on the mundane. Clinicians in recent decades have tended to focus on dramatic factors that may influence behavior in people with diabetes, such as depression, eating disorders, or fear of hypoglycemia. Certainly these conditions should be treated when clinically significant, but these may not be as common as clinicians believe, he said.

Old data suggested that 23%-45% of people with diabetes have depressive symptoms, but newer data suggest that only 5% of people with type 1 diabetes and 4% of those with type 2 diabetes meet the criteria for major depressive disorder, comparable to percentages in the general population, Dr. Polonsky said. Other data suggest that eating disorders may affect 10% of middle-aged adults with type 2 diabetes and 10% of young women with type 1 diabetes. "My hunch is that it’s not as big an issue as we think it is," he said. As many as 26% of people with type 2 diabetes and an unknown proportion of people with type 1 diabetes may fear hypoglycemia, "but I’m not sure we’re talking about big, dramatic fears," he said.

A bigger obstacle may be what Dr. Polonsky called "diabetes meh" – indifference or apathy, when patients just don’t’ really care. This can manifest in statements like, "I have more important things to worry about than diabetes," or "This disease will eventually get me no matter what I do," or "Changing how I eat is too much of a bother."

A related concept is "diabetes distress" – patients’ feelings that diabetes takes up too much mental and physical energy, that the disease controls their lives, and that they often fail with their diabetes regimen. "Should we rename this ‘diabetes fatigue’?" he asked. As many as 39% of patients with type 1 diabetes and 35% of patients with type 2 diabetes may suffer from diabetes distress and its sense of powerlessness, based on research reports.

Try focusing on conveying the "worthwhileness" of interventions, Dr. Polonsky suggested. One way to do that is by "making the invisible visible" by asking patients to check their blood glucose levels right before and right after taking a 45-minute walk each day for just 1 week. The results may surprise them in a motivating way.

3. Not helping patients choose actions that matter. Help patients understand which actions and strategies are more valuable than others. Think of it as looking for "the biggest bang for your buck," he said.

For most patients, that would be knowing their numbers (for glucose and HbA_1c), and taking the right medications. Those are more important than, for example, a deprivation strategy like choosing to eat two fewer tortillas per day, he said. Collaborate with patients to select actions that can make a real difference.

Dr. Polonsky reported having no relevant financial disclosures.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Getting patients to abandon unhealthful behaviors to better control diabetes can be hard, and sometimes the physician is the problem, according to behavioral psychologist William H. Polonsky, Ph.D.

"I think we’re a little too in love with the idea of behavior change," Dr. Polonsky said in the Richard R. Rubin Award lecture at the annual scientific sessions of the American Diabetes Association. "The major thing I’ve learned from my patients is we need to look in the mirror" to figure out why efforts at behavior change aren’t working.

Certainly, healthy behaviors are incredibly helpful. Patients who reported a high level of disease self-management were nearly 40% less likely to die over a 12-year period in a study of 340 patients with type 2 diabetes (Diabetes Care 2014; 37:1604-12). But when behavior change isn’t happening, think about three common mistakes that clinicians make in the process, said Dr. Polonsky, president of the nonprofit Behavioral Diabetes Institute, San Diego:

1. Not taking enough time. Don’t rush patients to change their unhealthful behavior too soon, Dr. Polonsky advised. Make certain your patient believes that the selected behavior change is worthwhile.

"I push my patients way too hard, and I push them when I shouldn’t be pushing them," he said. "It’s this urge we have because we care and probably because we’re overtrained as problem-solvers."

Many clinicians now employ motivational interviewing, for example, which is supposed to end with formulation of concrete, realistic goals for change. However, in reviewing four fairly large randomized, controlled trials published in 2010-2013 that compared motivational interviewing techniques to usual care in managing patients with diabetes, Dr. Polonsky found that the change in hemoglobin A_1c (HbA_1c) levels differed between groups by an average of only 0.1%. Two other studies found that using motivational interviewing was significantly less effective than diabetes education in changing HbA_1c levels.

Perhaps, the emphasis on formulating a concrete goal does not allow enough time to convince a patient that a specific behavior change would be worthwhile, he said.

2. Not focusing on the mundane. Clinicians in recent decades have tended to focus on dramatic factors that may influence behavior in people with diabetes, such as depression, eating disorders, or fear of hypoglycemia. Certainly these conditions should be treated when clinically significant, but these may not be as common as clinicians believe, he said.

Old data suggested that 23%-45% of people with diabetes have depressive symptoms, but newer data suggest that only 5% of people with type 1 diabetes and 4% of those with type 2 diabetes meet the criteria for major depressive disorder, comparable to percentages in the general population, Dr. Polonsky said. Other data suggest that eating disorders may affect 10% of middle-aged adults with type 2 diabetes and 10% of young women with type 1 diabetes. "My hunch is that it’s not as big an issue as we think it is," he said. As many as 26% of people with type 2 diabetes and an unknown proportion of people with type 1 diabetes may fear hypoglycemia, "but I’m not sure we’re talking about big, dramatic fears," he said.

A bigger obstacle may be what Dr. Polonsky called "diabetes meh" – indifference or apathy, when patients just don’t’ really care. This can manifest in statements like, "I have more important things to worry about than diabetes," or "This disease will eventually get me no matter what I do," or "Changing how I eat is too much of a bother."

A related concept is "diabetes distress" – patients’ feelings that diabetes takes up too much mental and physical energy, that the disease controls their lives, and that they often fail with their diabetes regimen. "Should we rename this ‘diabetes fatigue’?" he asked. As many as 39% of patients with type 1 diabetes and 35% of patients with type 2 diabetes may suffer from diabetes distress and its sense of powerlessness, based on research reports.

Try focusing on conveying the "worthwhileness" of interventions, Dr. Polonsky suggested. One way to do that is by "making the invisible visible" by asking patients to check their blood glucose levels right before and right after taking a 45-minute walk each day for just 1 week. The results may surprise them in a motivating way.

3. Not helping patients choose actions that matter. Help patients understand which actions and strategies are more valuable than others. Think of it as looking for "the biggest bang for your buck," he said.

For most patients, that would be knowing their numbers (for glucose and HbA_1c), and taking the right medications. Those are more important than, for example, a deprivation strategy like choosing to eat two fewer tortillas per day, he said. Collaborate with patients to select actions that can make a real difference.

Dr. Polonsky reported having no relevant financial disclosures.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Getting patients to abandon unhealthful behaviors to better control diabetes can be hard, and sometimes the physician is the problem, according to behavioral psychologist William H. Polonsky, Ph.D.

"I think we’re a little too in love with the idea of behavior change," Dr. Polonsky said in the Richard R. Rubin Award lecture at the annual scientific sessions of the American Diabetes Association. "The major thing I’ve learned from my patients is we need to look in the mirror" to figure out why efforts at behavior change aren’t working.

Certainly, healthy behaviors are incredibly helpful. Patients who reported a high level of disease self-management were nearly 40% less likely to die over a 12-year period in a study of 340 patients with type 2 diabetes (Diabetes Care 2014; 37:1604-12). But when behavior change isn’t happening, think about three common mistakes that clinicians make in the process, said Dr. Polonsky, president of the nonprofit Behavioral Diabetes Institute, San Diego:

1. Not taking enough time. Don’t rush patients to change their unhealthful behavior too soon, Dr. Polonsky advised. Make certain your patient believes that the selected behavior change is worthwhile.

"I push my patients way too hard, and I push them when I shouldn’t be pushing them," he said. "It’s this urge we have because we care and probably because we’re overtrained as problem-solvers."

Many clinicians now employ motivational interviewing, for example, which is supposed to end with formulation of concrete, realistic goals for change. However, in reviewing four fairly large randomized, controlled trials published in 2010-2013 that compared motivational interviewing techniques to usual care in managing patients with diabetes, Dr. Polonsky found that the change in hemoglobin A_1c (HbA_1c) levels differed between groups by an average of only 0.1%. Two other studies found that using motivational interviewing was significantly less effective than diabetes education in changing HbA_1c levels.

Perhaps, the emphasis on formulating a concrete goal does not allow enough time to convince a patient that a specific behavior change would be worthwhile, he said.

2. Not focusing on the mundane. Clinicians in recent decades have tended to focus on dramatic factors that may influence behavior in people with diabetes, such as depression, eating disorders, or fear of hypoglycemia. Certainly these conditions should be treated when clinically significant, but these may not be as common as clinicians believe, he said.

Old data suggested that 23%-45% of people with diabetes have depressive symptoms, but newer data suggest that only 5% of people with type 1 diabetes and 4% of those with type 2 diabetes meet the criteria for major depressive disorder, comparable to percentages in the general population, Dr. Polonsky said. Other data suggest that eating disorders may affect 10% of middle-aged adults with type 2 diabetes and 10% of young women with type 1 diabetes. "My hunch is that it’s not as big an issue as we think it is," he said. As many as 26% of people with type 2 diabetes and an unknown proportion of people with type 1 diabetes may fear hypoglycemia, "but I’m not sure we’re talking about big, dramatic fears," he said.

A bigger obstacle may be what Dr. Polonsky called "diabetes meh" – indifference or apathy, when patients just don’t’ really care. This can manifest in statements like, "I have more important things to worry about than diabetes," or "This disease will eventually get me no matter what I do," or "Changing how I eat is too much of a bother."

A related concept is "diabetes distress" – patients’ feelings that diabetes takes up too much mental and physical energy, that the disease controls their lives, and that they often fail with their diabetes regimen. "Should we rename this ‘diabetes fatigue’?" he asked. As many as 39% of patients with type 1 diabetes and 35% of patients with type 2 diabetes may suffer from diabetes distress and its sense of powerlessness, based on research reports.

Try focusing on conveying the "worthwhileness" of interventions, Dr. Polonsky suggested. One way to do that is by "making the invisible visible" by asking patients to check their blood glucose levels right before and right after taking a 45-minute walk each day for just 1 week. The results may surprise them in a motivating way.

3. Not helping patients choose actions that matter. Help patients understand which actions and strategies are more valuable than others. Think of it as looking for "the biggest bang for your buck," he said.

For most patients, that would be knowing their numbers (for glucose and HbA_1c), and taking the right medications. Those are more important than, for example, a deprivation strategy like choosing to eat two fewer tortillas per day, he said. Collaborate with patients to select actions that can make a real difference.

Dr. Polonsky reported having no relevant financial disclosures.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Hypoglycemic episodes were common and largely unnoticed after bariatric surgery, a controlled study of 45 patients found.

During a 3-day period of "normal living," symptomatic hypoglycemias occurred in 22% of 15 patients after gastric bypass surgery, 20% of 15 patients after biliopancreatic diversion with duodenal switch, and in none of 15 obese, nondiabetic control patients matched to the surgical patients by body mass index.

Continuous glucose monitoring showed that patients in both postsurgery groups spent significant amounts of time in hypoglycemia, Dr. Niclas Abrahamsson and his associates reported at the annual scientific session of the American Diabetes Association.

After gastric bypass, patients averaged 42 minutes per day with glucose levels lower than 3.3 mmol/L and 21 minutes per day with levels lower than 2.8 mmol/L. After duodenal switch surgery, patients averaged 85 minutes per day with glucose levels lower than 3.3 mmol/L and 39 minutes per day with levels lower than 2.8 mmol/L. No patients in the control group had glucose levels that low, reported Dr. Abrahamsson of the University of Uppsala, Sweden.

"We were very surprised that they had so many hypoglycemic episodes, especially since the controls had none," he said. Patients were unaware of approximately 80% of the hypoglycemic episodes, he added.

"The clinical significance should be that one should be alert to any hypoglycemia symptoms," Dr. Abrahamsson said.

Patients in the post–duodenal switch group had the lowest mean glucose level (4.6 mmol/L) and mean hemoglobin A_1c level (29 mmol/mol), compared with the post–gastric bypass group (mean glucose 5.3 mmol/L and HbA_1c 36 mmol/mol) and the control group (mean glucose 5.9 mmol/L and HbA_1c 38 mmol/mol).

Glucose curves on continuous monitoring were more variable in the post–gastric bypass group, compared with controls, and less variable in the post–duodenal switch group, compared with controls. That difference between the two surgical groups probably relates to the different glucose absorption capabilities after surgery, he suggested.

Dr. Abrahamsson has been a speaker for Eli Lilly and Sanofi and has held stock in AstraZeneca.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Hypoglycemic episodes were common and largely unnoticed after bariatric surgery, a controlled study of 45 patients found.

During a 3-day period of "normal living," symptomatic hypoglycemias occurred in 22% of 15 patients after gastric bypass surgery, 20% of 15 patients after biliopancreatic diversion with duodenal switch, and in none of 15 obese, nondiabetic control patients matched to the surgical patients by body mass index.

Continuous glucose monitoring showed that patients in both postsurgery groups spent significant amounts of time in hypoglycemia, Dr. Niclas Abrahamsson and his associates reported at the annual scientific session of the American Diabetes Association.

After gastric bypass, patients averaged 42 minutes per day with glucose levels lower than 3.3 mmol/L and 21 minutes per day with levels lower than 2.8 mmol/L. After duodenal switch surgery, patients averaged 85 minutes per day with glucose levels lower than 3.3 mmol/L and 39 minutes per day with levels lower than 2.8 mmol/L. No patients in the control group had glucose levels that low, reported Dr. Abrahamsson of the University of Uppsala, Sweden.

"We were very surprised that they had so many hypoglycemic episodes, especially since the controls had none," he said. Patients were unaware of approximately 80% of the hypoglycemic episodes, he added.

"The clinical significance should be that one should be alert to any hypoglycemia symptoms," Dr. Abrahamsson said.

Patients in the post–duodenal switch group had the lowest mean glucose level (4.6 mmol/L) and mean hemoglobin A_1c level (29 mmol/mol), compared with the post–gastric bypass group (mean glucose 5.3 mmol/L and HbA_1c 36 mmol/mol) and the control group (mean glucose 5.9 mmol/L and HbA_1c 38 mmol/mol).

Glucose curves on continuous monitoring were more variable in the post–gastric bypass group, compared with controls, and less variable in the post–duodenal switch group, compared with controls. That difference between the two surgical groups probably relates to the different glucose absorption capabilities after surgery, he suggested.

Dr. Abrahamsson has been a speaker for Eli Lilly and Sanofi and has held stock in AstraZeneca.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Hypoglycemic episodes were common and largely unnoticed after bariatric surgery, a controlled study of 45 patients found.

During a 3-day period of "normal living," symptomatic hypoglycemias occurred in 22% of 15 patients after gastric bypass surgery, 20% of 15 patients after biliopancreatic diversion with duodenal switch, and in none of 15 obese, nondiabetic control patients matched to the surgical patients by body mass index.

Continuous glucose monitoring showed that patients in both postsurgery groups spent significant amounts of time in hypoglycemia, Dr. Niclas Abrahamsson and his associates reported at the annual scientific session of the American Diabetes Association.

After gastric bypass, patients averaged 42 minutes per day with glucose levels lower than 3.3 mmol/L and 21 minutes per day with levels lower than 2.8 mmol/L. After duodenal switch surgery, patients averaged 85 minutes per day with glucose levels lower than 3.3 mmol/L and 39 minutes per day with levels lower than 2.8 mmol/L. No patients in the control group had glucose levels that low, reported Dr. Abrahamsson of the University of Uppsala, Sweden.

"We were very surprised that they had so many hypoglycemic episodes, especially since the controls had none," he said. Patients were unaware of approximately 80% of the hypoglycemic episodes, he added.

"The clinical significance should be that one should be alert to any hypoglycemia symptoms," Dr. Abrahamsson said.

Patients in the post–duodenal switch group had the lowest mean glucose level (4.6 mmol/L) and mean hemoglobin A_1c level (29 mmol/mol), compared with the post–gastric bypass group (mean glucose 5.3 mmol/L and HbA_1c 36 mmol/mol) and the control group (mean glucose 5.9 mmol/L and HbA_1c 38 mmol/mol).

Glucose curves on continuous monitoring were more variable in the post–gastric bypass group, compared with controls, and less variable in the post–duodenal switch group, compared with controls. That difference between the two surgical groups probably relates to the different glucose absorption capabilities after surgery, he suggested.

Dr. Abrahamsson has been a speaker for Eli Lilly and Sanofi and has held stock in AstraZeneca.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Adults who started taking sulfonylurea drugs for diabetes were 9% more likely to develop fractures within 5 years, and patients starting thiazolidinediones were 9% more likely to have fractures, compared with patients on other medications in a retrospective study of 99,892 adults.

Those significantly increased risks emerged after adjusting for the effects of multiple factors including age, gender, region, medical conditions, and concomitant medications, Sandhya Mehta, Ph.D. and her associates reported at the annual scientific sessions of the American Diabetes Association. Patients had no prior history of fracture.

The longitudinal retrospective analysis of a large administrative claims database found a 10% incidence of fracture on sulfonylureas and 11% on thiazolidinediones, compared with 7% on biguanides (metformin), 8% on dipeptidyl peptidase-4 (DPP-4) inhibitors, 11% on meglitinide analogues, and 6% on incretin mimetic agents. After adjusting for the confounders, the hazard ratios for each of the other three types of drugs hovered around 1 and were not statistically significantly different but increased to 1.09 for sulfonylureas and 1.4 for thiazolidinediones, reported Dr. Mehta of Inovalon Inc. in Bowie, Md., a health care data analytics company.

In the cohort as a whole, 7% of patients developed fractures. Roughly 15% of patients started sulfonylureas, 3% took thiazolidinediones, 78% were on metformin, 3% took DPP-4 inhibitors, 1% were on incretin mimetics, and 1% took meglitinides.

Previous reports have shown an increased risk of fracture in patients on thiazolidinediones, compared with those on metformin, and the current results support the hypothesis that thiazolidinediones decrease bone mineral density, stimulate adipocyte and osteoclast differentiation, and inhibit osteoblast differentiation to make fracture more likely, Dr. Mehta said.

The association between sulfonylureas and increased fracture risk appears to be new, however, and deserves further study, she added.

The study did not look at the effects of combination drug therapy.

Data came from the Medical Outcomes Research for Effectiveness and Economics (MORE) Registry, which is owned by Inovalon, drawing from the years 2008-2012.

Dr. Mehta reported having no financial disclosures.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Adults who started taking sulfonylurea drugs for diabetes were 9% more likely to develop fractures within 5 years, and patients starting thiazolidinediones were 9% more likely to have fractures, compared with patients on other medications in a retrospective study of 99,892 adults.

Those significantly increased risks emerged after adjusting for the effects of multiple factors including age, gender, region, medical conditions, and concomitant medications, Sandhya Mehta, Ph.D. and her associates reported at the annual scientific sessions of the American Diabetes Association. Patients had no prior history of fracture.

The longitudinal retrospective analysis of a large administrative claims database found a 10% incidence of fracture on sulfonylureas and 11% on thiazolidinediones, compared with 7% on biguanides (metformin), 8% on dipeptidyl peptidase-4 (DPP-4) inhibitors, 11% on meglitinide analogues, and 6% on incretin mimetic agents. After adjusting for the confounders, the hazard ratios for each of the other three types of drugs hovered around 1 and were not statistically significantly different but increased to 1.09 for sulfonylureas and 1.4 for thiazolidinediones, reported Dr. Mehta of Inovalon Inc. in Bowie, Md., a health care data analytics company.

In the cohort as a whole, 7% of patients developed fractures. Roughly 15% of patients started sulfonylureas, 3% took thiazolidinediones, 78% were on metformin, 3% took DPP-4 inhibitors, 1% were on incretin mimetics, and 1% took meglitinides.

Previous reports have shown an increased risk of fracture in patients on thiazolidinediones, compared with those on metformin, and the current results support the hypothesis that thiazolidinediones decrease bone mineral density, stimulate adipocyte and osteoclast differentiation, and inhibit osteoblast differentiation to make fracture more likely, Dr. Mehta said.

The association between sulfonylureas and increased fracture risk appears to be new, however, and deserves further study, she added.

The study did not look at the effects of combination drug therapy.

Data came from the Medical Outcomes Research for Effectiveness and Economics (MORE) Registry, which is owned by Inovalon, drawing from the years 2008-2012.

Dr. Mehta reported having no financial disclosures.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Adults who started taking sulfonylurea drugs for diabetes were 9% more likely to develop fractures within 5 years, and patients starting thiazolidinediones were 9% more likely to have fractures, compared with patients on other medications in a retrospective study of 99,892 adults.

Those significantly increased risks emerged after adjusting for the effects of multiple factors including age, gender, region, medical conditions, and concomitant medications, Sandhya Mehta, Ph.D. and her associates reported at the annual scientific sessions of the American Diabetes Association. Patients had no prior history of fracture.

The longitudinal retrospective analysis of a large administrative claims database found a 10% incidence of fracture on sulfonylureas and 11% on thiazolidinediones, compared with 7% on biguanides (metformin), 8% on dipeptidyl peptidase-4 (DPP-4) inhibitors, 11% on meglitinide analogues, and 6% on incretin mimetic agents. After adjusting for the confounders, the hazard ratios for each of the other three types of drugs hovered around 1 and were not statistically significantly different but increased to 1.09 for sulfonylureas and 1.4 for thiazolidinediones, reported Dr. Mehta of Inovalon Inc. in Bowie, Md., a health care data analytics company.

In the cohort as a whole, 7% of patients developed fractures. Roughly 15% of patients started sulfonylureas, 3% took thiazolidinediones, 78% were on metformin, 3% took DPP-4 inhibitors, 1% were on incretin mimetics, and 1% took meglitinides.

Previous reports have shown an increased risk of fracture in patients on thiazolidinediones, compared with those on metformin, and the current results support the hypothesis that thiazolidinediones decrease bone mineral density, stimulate adipocyte and osteoclast differentiation, and inhibit osteoblast differentiation to make fracture more likely, Dr. Mehta said.

The association between sulfonylureas and increased fracture risk appears to be new, however, and deserves further study, she added.

The study did not look at the effects of combination drug therapy.

Data came from the Medical Outcomes Research for Effectiveness and Economics (MORE) Registry, which is owned by Inovalon, drawing from the years 2008-2012.

Dr. Mehta reported having no financial disclosures.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Physicians and advanced practice health care providers saw comparable types of patients with diabetes in the Veterans Affairs health care system and helped patients achieve comparable changes in hemoglobin A1C levels, a study of 19,238 health care providers found.

"At the time of diagnosis of diabetes and during 4 years of follow-up care, nurse practitioners and physician assistants have patient populations that are comparable to those of physicians, and their diabetes management is as good as that provided by physicians," Dr. Lawrence S. Phillips said at the annual scientific sessions of the American Diabetes Association.

Diabetes affects more than 20% of patients in the VA system, which is the largest employer of nurse practitioners and physician assistants in the country, he said. Dr. Phillips and his associates analyzed VA data to identify patients who saw a physician, nurse practitioner, or physician assistant for more than 50% of health care visits since a new diagnosis of diabetes in 2008-2012.

Average A1C levels at the time of diabetes diagnosis, at the initiation of medications, and during follow-up years did not differ significantly between provider groups after patient characteristics were adjusted for, reported Dr. Phillips, professor of medicine at Emory University, Atlanta. Baseline A1C levels were 7.18%, 7.14%, and 7.20% with nurse practitioners, physician assistants, and physicians, respectively, and dropped to 6.75%, 6.75%, and 6.78% by 2012.

Patient demographics were similar between provider groups. Their mean age was roughly 69 years, about 95% were male, and their mean body mass index was about 32.4 kg/m². The cohort seen by physicians had a smaller percentage of white patients (77%) than did the cohort seen by physician assistants (81%).

Patient comorbidity levels, as measured by the Charlson Comorbidity Index, were the same in each provider group (a score of 1 in each group). The proportion of patients on any diabetes therapy was similar between groups: 81% with physicians, 80% with nurse practitioners, and 79% with physician assistants.

A significantly smaller proportion of patients seeing physician assistants were on insulin (12%), compared with those seeing physicians (15%) or nurse practitioners (14%), Dr. Phillips reported. Physician assistants also referred a significantly smaller proportion of patients to diabetes specialists (5%), compared with physicians (8%) or nurse practitioners (7%).

The findings surprised the investigators and suggested that the VA model of using advanced practice providers might be broadly useful to help the U.S. meet demand for primary care providers, he said.

Dr. Phillips reported having financial associations with multiple pharmaceutical companies but none related to this study.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Physicians and advanced practice health care providers saw comparable types of patients with diabetes in the Veterans Affairs health care system and helped patients achieve comparable changes in hemoglobin A1C levels, a study of 19,238 health care providers found.

"At the time of diagnosis of diabetes and during 4 years of follow-up care, nurse practitioners and physician assistants have patient populations that are comparable to those of physicians, and their diabetes management is as good as that provided by physicians," Dr. Lawrence S. Phillips said at the annual scientific sessions of the American Diabetes Association.

Diabetes affects more than 20% of patients in the VA system, which is the largest employer of nurse practitioners and physician assistants in the country, he said. Dr. Phillips and his associates analyzed VA data to identify patients who saw a physician, nurse practitioner, or physician assistant for more than 50% of health care visits since a new diagnosis of diabetes in 2008-2012.

Average A1C levels at the time of diabetes diagnosis, at the initiation of medications, and during follow-up years did not differ significantly between provider groups after patient characteristics were adjusted for, reported Dr. Phillips, professor of medicine at Emory University, Atlanta. Baseline A1C levels were 7.18%, 7.14%, and 7.20% with nurse practitioners, physician assistants, and physicians, respectively, and dropped to 6.75%, 6.75%, and 6.78% by 2012.

Patient demographics were similar between provider groups. Their mean age was roughly 69 years, about 95% were male, and their mean body mass index was about 32.4 kg/m². The cohort seen by physicians had a smaller percentage of white patients (77%) than did the cohort seen by physician assistants (81%).

Patient comorbidity levels, as measured by the Charlson Comorbidity Index, were the same in each provider group (a score of 1 in each group). The proportion of patients on any diabetes therapy was similar between groups: 81% with physicians, 80% with nurse practitioners, and 79% with physician assistants.

A significantly smaller proportion of patients seeing physician assistants were on insulin (12%), compared with those seeing physicians (15%) or nurse practitioners (14%), Dr. Phillips reported. Physician assistants also referred a significantly smaller proportion of patients to diabetes specialists (5%), compared with physicians (8%) or nurse practitioners (7%).

The findings surprised the investigators and suggested that the VA model of using advanced practice providers might be broadly useful to help the U.S. meet demand for primary care providers, he said.

Dr. Phillips reported having financial associations with multiple pharmaceutical companies but none related to this study.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Physicians and advanced practice health care providers saw comparable types of patients with diabetes in the Veterans Affairs health care system and helped patients achieve comparable changes in hemoglobin A1C levels, a study of 19,238 health care providers found.

"At the time of diagnosis of diabetes and during 4 years of follow-up care, nurse practitioners and physician assistants have patient populations that are comparable to those of physicians, and their diabetes management is as good as that provided by physicians," Dr. Lawrence S. Phillips said at the annual scientific sessions of the American Diabetes Association.

Diabetes affects more than 20% of patients in the VA system, which is the largest employer of nurse practitioners and physician assistants in the country, he said. Dr. Phillips and his associates analyzed VA data to identify patients who saw a physician, nurse practitioner, or physician assistant for more than 50% of health care visits since a new diagnosis of diabetes in 2008-2012.

Average A1C levels at the time of diabetes diagnosis, at the initiation of medications, and during follow-up years did not differ significantly between provider groups after patient characteristics were adjusted for, reported Dr. Phillips, professor of medicine at Emory University, Atlanta. Baseline A1C levels were 7.18%, 7.14%, and 7.20% with nurse practitioners, physician assistants, and physicians, respectively, and dropped to 6.75%, 6.75%, and 6.78% by 2012.

Patient demographics were similar between provider groups. Their mean age was roughly 69 years, about 95% were male, and their mean body mass index was about 32.4 kg/m². The cohort seen by physicians had a smaller percentage of white patients (77%) than did the cohort seen by physician assistants (81%).

Patient comorbidity levels, as measured by the Charlson Comorbidity Index, were the same in each provider group (a score of 1 in each group). The proportion of patients on any diabetes therapy was similar between groups: 81% with physicians, 80% with nurse practitioners, and 79% with physician assistants.

A significantly smaller proportion of patients seeing physician assistants were on insulin (12%), compared with those seeing physicians (15%) or nurse practitioners (14%), Dr. Phillips reported. Physician assistants also referred a significantly smaller proportion of patients to diabetes specialists (5%), compared with physicians (8%) or nurse practitioners (7%).

The findings surprised the investigators and suggested that the VA model of using advanced practice providers might be broadly useful to help the U.S. meet demand for primary care providers, he said.

Dr. Phillips reported having financial associations with multiple pharmaceutical companies but none related to this study.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Physicians and midlevel health care providers saw comparable types of patients with diabetes in the Veterans Affairs health care system and helped patients achieve comparable changes in hemoglobin A_1c levels, a study of 19,238 health care providers found.

"At the time of diagnosis of diabetes and during 4 years of follow-up care, nurse practitioners and physician assistants have patient populations that are comparable to those of physicians, and their diabetes management is as good as that provided by physicians," Dr. Lawrence S. Phillips said at the annual scientific sessions of the American Diabetes Association.

Diabetes affects more than 20% of patients in the VA system, which is the largest employer of nurse practitioners and physician assistants in the country, he said. Dr. Phillips and his associates analyzed VA data to identify patients who saw a physician, nurse practitioner, or physician assistant for more than 50% of health care visits since a new diagnosis of diabetes in 2008-2012.

Average HbA_1c levels at the time of diabetes diagnosis, at the initiation of medications, and during follow-up years did not differ significantly between provider groups after patient characteristics were adjusted for, reported Dr. Phillips, professor of medicine at Emory University, Atlanta. Baseline HbA_1c levels were 7.18%, 7.14%, and 7.20% with nurse practitioners, physician assistants, and physicians, respectively, and dropped to 6.75%, 6.75%, and 6.78% by 2012.

Patient demographics were similar between provider groups. Their mean age was roughly 69 years, about 95% were male, and their mean body mass index was about 32.4 kg/m². The cohort seen by physicians had a smaller percentage of white patients (77%) than did the cohort seen by physician assistants (81%).

Patient comorbidity levels, as measured by the Charlson Comorbidity Index, were the same in each provider group (a score of 1 in each group). The proportion of patients on any diabetes therapy was similar between groups: 81% with physicians, 80% with nurse practitioners, and 79% with physician assistants.

A significantly smaller proportion of patients seeing physician assistants were on insulin (12%), compared with those seeing physicians (15%) or nurse practitioners (14%), Dr. Phillips reported. Physician assistants also referred a significantly smaller proportion of patients to diabetes specialists (5%), compared with physicians (8%) or nurse practitioners (7%).

The findings surprised the investigators and suggested that the VA model of using midlevel providers might be broadly useful to help the U.S. meet demand for primary care providers, he said.

Dr. Phillips reported having financial associations with multiple pharmaceutical companies but none related to this study.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Physicians and midlevel health care providers saw comparable types of patients with diabetes in the Veterans Affairs health care system and helped patients achieve comparable changes in hemoglobin A_1c levels, a study of 19,238 health care providers found.

"At the time of diagnosis of diabetes and during 4 years of follow-up care, nurse practitioners and physician assistants have patient populations that are comparable to those of physicians, and their diabetes management is as good as that provided by physicians," Dr. Lawrence S. Phillips said at the annual scientific sessions of the American Diabetes Association.

Diabetes affects more than 20% of patients in the VA system, which is the largest employer of nurse practitioners and physician assistants in the country, he said. Dr. Phillips and his associates analyzed VA data to identify patients who saw a physician, nurse practitioner, or physician assistant for more than 50% of health care visits since a new diagnosis of diabetes in 2008-2012.

Average HbA_1c levels at the time of diabetes diagnosis, at the initiation of medications, and during follow-up years did not differ significantly between provider groups after patient characteristics were adjusted for, reported Dr. Phillips, professor of medicine at Emory University, Atlanta. Baseline HbA_1c levels were 7.18%, 7.14%, and 7.20% with nurse practitioners, physician assistants, and physicians, respectively, and dropped to 6.75%, 6.75%, and 6.78% by 2012.

Patient demographics were similar between provider groups. Their mean age was roughly 69 years, about 95% were male, and their mean body mass index was about 32.4 kg/m². The cohort seen by physicians had a smaller percentage of white patients (77%) than did the cohort seen by physician assistants (81%).

Patient comorbidity levels, as measured by the Charlson Comorbidity Index, were the same in each provider group (a score of 1 in each group). The proportion of patients on any diabetes therapy was similar between groups: 81% with physicians, 80% with nurse practitioners, and 79% with physician assistants.

A significantly smaller proportion of patients seeing physician assistants were on insulin (12%), compared with those seeing physicians (15%) or nurse practitioners (14%), Dr. Phillips reported. Physician assistants also referred a significantly smaller proportion of patients to diabetes specialists (5%), compared with physicians (8%) or nurse practitioners (7%).

The findings surprised the investigators and suggested that the VA model of using midlevel providers might be broadly useful to help the U.S. meet demand for primary care providers, he said.

Dr. Phillips reported having financial associations with multiple pharmaceutical companies but none related to this study.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Physicians and midlevel health care providers saw comparable types of patients with diabetes in the Veterans Affairs health care system and helped patients achieve comparable changes in hemoglobin A_1c levels, a study of 19,238 health care providers found.

"At the time of diagnosis of diabetes and during 4 years of follow-up care, nurse practitioners and physician assistants have patient populations that are comparable to those of physicians, and their diabetes management is as good as that provided by physicians," Dr. Lawrence S. Phillips said at the annual scientific sessions of the American Diabetes Association.

Diabetes affects more than 20% of patients in the VA system, which is the largest employer of nurse practitioners and physician assistants in the country, he said. Dr. Phillips and his associates analyzed VA data to identify patients who saw a physician, nurse practitioner, or physician assistant for more than 50% of health care visits since a new diagnosis of diabetes in 2008-2012.

Average HbA_1c levels at the time of diabetes diagnosis, at the initiation of medications, and during follow-up years did not differ significantly between provider groups after patient characteristics were adjusted for, reported Dr. Phillips, professor of medicine at Emory University, Atlanta. Baseline HbA_1c levels were 7.18%, 7.14%, and 7.20% with nurse practitioners, physician assistants, and physicians, respectively, and dropped to 6.75%, 6.75%, and 6.78% by 2012.

Patient demographics were similar between provider groups. Their mean age was roughly 69 years, about 95% were male, and their mean body mass index was about 32.4 kg/m². The cohort seen by physicians had a smaller percentage of white patients (77%) than did the cohort seen by physician assistants (81%).

Patient comorbidity levels, as measured by the Charlson Comorbidity Index, were the same in each provider group (a score of 1 in each group). The proportion of patients on any diabetes therapy was similar between groups: 81% with physicians, 80% with nurse practitioners, and 79% with physician assistants.

A significantly smaller proportion of patients seeing physician assistants were on insulin (12%), compared with those seeing physicians (15%) or nurse practitioners (14%), Dr. Phillips reported. Physician assistants also referred a significantly smaller proportion of patients to diabetes specialists (5%), compared with physicians (8%) or nurse practitioners (7%).

The findings surprised the investigators and suggested that the VA model of using midlevel providers might be broadly useful to help the U.S. meet demand for primary care providers, he said.

Dr. Phillips reported having financial associations with multiple pharmaceutical companies but none related to this study.

sboschert@frontlinemedcom.com

On Twitter @sherryboschert

SAN FRANCISCO – Veterans who participated in a lifestyle change program to reduce weight and increase exercise experienced a 17% lower total incidence of cardiovascular disease over 5 years, results from a large analysis demonstrated.

"Two-thirds of U.S. adults are overweight or obese, yet we know little about how to bring help to them," Sandra L. Jackson, Ph.D., said at the annual scientific sessions of the American Diabetes Association.

Lifestyle change is a recommended strategy for prevention of cardiovascular disease, but the evidence of effect of lifestyle change programs on CVD "is mixed, and few studies have had sufficient sample size to study CVD incidence as compared to change in CVD risk factors. In addition, we know little about what lifestyle change programs can achieve in health care settings, where participants are patients and providers recommend a change in lifestyle."

For the current study, she and her associates used data from the U.S. Department of Veterans Affairs’ national weight management program, MOVE! The program has enrolled more than 400,000 veterans since 2005, making it the largest lifestyle change program in the United States. Implementation varies across VA facilities, but most use a standard 10-session core curriculum, with topics that include reading food labels, reducing fat intake, evaluating portion sizes, walking with a pedometer, and setting physical activity goals.

"The program is based on principles of motivational interviewing, so VA patients are encouraged to set their own goals for physical activity and weight loss, rather than follow a set program of prescribed goals," said Dr. Jackson, who conducted the study during her doctoral work in the Nutrition and Health Sciences Program at Emory University, Atlanta. "Nearly three-quarters of patients in the VA health system are overweight or obese and in theory, could benefit from MOVE!"

The researchers evaluated MOVE! participants with data from the National VA Informatics and Computing Infrastructure, which collects data from medical records, including demographics, outpatient visits, inpatient records, diagnoses, procedures, and prescriptions, and also features statistical tools. The study included 1,463,003 eligible patients who did not have CVD at baseline. Their mean age was 52 years, 89% were men, and their mean BMI was 33.6 kg/m². The average follow-up was about 5 years. Of the 1,463,003 patients, 169,248 were MOVE! participants and 1,293,755 were not.

Dr. Jackson reported that MOVE! participants who engaged in at least 8 sessions over 6 months lost about 2.5% of their body weight, and mostly sustained the loss over 3 years. The less active MOVE! participants lost a little over 0.5% of their body weight and maintained the loss over 3 years, while the eligible nonparticipants gained about 0.5% of body weight over the 3 years.

"Even with the substantial differences between the participant and nonparticipant groups, direct comparisons of the risk factors are hard to interpret," Dr. Jackson noted. For this reason, she and her associates used least-square means to compute averages, and adjusted for differences in baseline CVD risk factors, as well as differences in baseline BMI, age, gender, and race/ethnicity between the populations. They also stratified the analyses by diabetes status.

Systolic BP was a little lower among MOVE! participants than among nonparticipants, particularly at 6 and 12 months, "although the effect disappeared over 24 and 36 months," Dr. Jackson said. HDL cholesterol was a little higher among participants, compared with nonparticipants, while non-HDL cholesterol and plasma glucose levels were a little lower among participants, compared with nonparticipants.

Cox proportional hazards modeling revealed that there was a 17% lower incidence of CVD among participants, compared with nonparticipants (hazard ratio, 0.83). "When we looked separately at coronary artery disease, stroke, peripheral vascular disease, and heart failure, the effects remained significant," she said. These models were also adjusted for age, gender, BMI, smoking status, and a propensity score indicating likelihood of MOVE! participation, as well as other demographic and clinical characteristics.

A subgroup analysis revealed that the effect of MOVE! participation was stronger for men than for women (HR 0.83 vs. HR 0.93, respectively; P = .001).

"The VA population does have substantially fewer women than men, and women in the VA tend to be younger and may be at lower CVD risk than men," Dr. Jackson noted. "We also saw a stronger effect of participation among VA patients without diabetes, compared with those with diabetes, and we saw a stronger effect of participation among current smokers compared with nonsmokers."

Dr. Jackson said that she had no relevant financial conflicts to disclose.

dbrunk@frontlinemedcom.com

On Twitter @dougbrunk 

SAN FRANCISCO – Veterans who participated in a lifestyle change program to reduce weight and increase exercise experienced a 17% lower total incidence of cardiovascular disease over 5 years, results from a large analysis demonstrated.

"Two-thirds of U.S. adults are overweight or obese, yet we know little about how to bring help to them," Sandra L. Jackson, Ph.D., said at the annual scientific sessions of the American Diabetes Association.

Lifestyle change is a recommended strategy for prevention of cardiovascular disease, but the evidence of effect of lifestyle change programs on CVD "is mixed, and few studies have had sufficient sample size to study CVD incidence as compared to change in CVD risk factors. In addition, we know little about what lifestyle change programs can achieve in health care settings, where participants are patients and providers recommend a change in lifestyle."

For the current study, she and her associates used data from the U.S. Department of Veterans Affairs’ national weight management program, MOVE! The program has enrolled more than 400,000 veterans since 2005, making it the largest lifestyle change program in the United States. Implementation varies across VA facilities, but most use a standard 10-session core curriculum, with topics that include reading food labels, reducing fat intake, evaluating portion sizes, walking with a pedometer, and setting physical activity goals.

"The program is based on principles of motivational interviewing, so VA patients are encouraged to set their own goals for physical activity and weight loss, rather than follow a set program of prescribed goals," said Dr. Jackson, who conducted the study during her doctoral work in the Nutrition and Health Sciences Program at Emory University, Atlanta. "Nearly three-quarters of patients in the VA health system are overweight or obese and in theory, could benefit from MOVE!"

The researchers evaluated MOVE! participants with data from the National VA Informatics and Computing Infrastructure, which collects data from medical records, including demographics, outpatient visits, inpatient records, diagnoses, procedures, and prescriptions, and also features statistical tools. The study included 1,463,003 eligible patients who did not have CVD at baseline. Their mean age was 52 years, 89% were men, and their mean BMI was 33.6 kg/m². The average follow-up was about 5 years. Of the 1,463,003 patients, 169,248 were MOVE! participants and 1,293,755 were not.

Dr. Jackson reported that MOVE! participants who engaged in at least 8 sessions over 6 months lost about 2.5% of their body weight, and mostly sustained the loss over 3 years. The less active MOVE! participants lost a little over 0.5% of their body weight and maintained the loss over 3 years, while the eligible nonparticipants gained about 0.5% of body weight over the 3 years.

"Even with the substantial differences between the participant and nonparticipant groups, direct comparisons of the risk factors are hard to interpret," Dr. Jackson noted. For this reason, she and her associates used least-square means to compute averages, and adjusted for differences in baseline CVD risk factors, as well as differences in baseline BMI, age, gender, and race/ethnicity between the populations. They also stratified the analyses by diabetes status.

Systolic BP was a little lower among MOVE! participants than among nonparticipants, particularly at 6 and 12 months, "although the effect disappeared over 24 and 36 months," Dr. Jackson said. HDL cholesterol was a little higher among participants, compared with nonparticipants, while non-HDL cholesterol and plasma glucose levels were a little lower among participants, compared with nonparticipants.

Cox proportional hazards modeling revealed that there was a 17% lower incidence of CVD among participants, compared with nonparticipants (hazard ratio, 0.83). "When we looked separately at coronary artery disease, stroke, peripheral vascular disease, and heart failure, the effects remained significant," she said. These models were also adjusted for age, gender, BMI, smoking status, and a propensity score indicating likelihood of MOVE! participation, as well as other demographic and clinical characteristics.

A subgroup analysis revealed that the effect of MOVE! participation was stronger for men than for women (HR 0.83 vs. HR 0.93, respectively; P = .001).

"The VA population does have substantially fewer women than men, and women in the VA tend to be younger and may be at lower CVD risk than men," Dr. Jackson noted. "We also saw a stronger effect of participation among VA patients without diabetes, compared with those with diabetes, and we saw a stronger effect of participation among current smokers compared with nonsmokers."

Dr. Jackson said that she had no relevant financial conflicts to disclose.

dbrunk@frontlinemedcom.com

On Twitter @dougbrunk 

SAN FRANCISCO – Veterans who participated in a lifestyle change program to reduce weight and increase exercise experienced a 17% lower total incidence of cardiovascular disease over 5 years, results from a large analysis demonstrated.

"Two-thirds of U.S. adults are overweight or obese, yet we know little about how to bring help to them," Sandra L. Jackson, Ph.D., said at the annual scientific sessions of the American Diabetes Association.

Lifestyle change is a recommended strategy for prevention of cardiovascular disease, but the evidence of effect of lifestyle change programs on CVD "is mixed, and few studies have had sufficient sample size to study CVD incidence as compared to change in CVD risk factors. In addition, we know little about what lifestyle change programs can achieve in health care settings, where participants are patients and providers recommend a change in lifestyle."

For the current study, she and her associates used data from the U.S. Department of Veterans Affairs’ national weight management program, MOVE! The program has enrolled more than 400,000 veterans since 2005, making it the largest lifestyle change program in the United States. Implementation varies across VA facilities, but most use a standard 10-session core curriculum, with topics that include reading food labels, reducing fat intake, evaluating portion sizes, walking with a pedometer, and setting physical activity goals.

"The program is based on principles of motivational interviewing, so VA patients are encouraged to set their own goals for physical activity and weight loss, rather than follow a set program of prescribed goals," said Dr. Jackson, who conducted the study during her doctoral work in the Nutrition and Health Sciences Program at Emory University, Atlanta. "Nearly three-quarters of patients in the VA health system are overweight or obese and in theory, could benefit from MOVE!"

The researchers evaluated MOVE! participants with data from the National VA Informatics and Computing Infrastructure, which collects data from medical records, including demographics, outpatient visits, inpatient records, diagnoses, procedures, and prescriptions, and also features statistical tools. The study included 1,463,003 eligible patients who did not have CVD at baseline. Their mean age was 52 years, 89% were men, and their mean BMI was 33.6 kg/m². The average follow-up was about 5 years. Of the 1,463,003 patients, 169,248 were MOVE! participants and 1,293,755 were not.

Dr. Jackson reported that MOVE! participants who engaged in at least 8 sessions over 6 months lost about 2.5% of their body weight, and mostly sustained the loss over 3 years. The less active MOVE! participants lost a little over 0.5% of their body weight and maintained the loss over 3 years, while the eligible nonparticipants gained about 0.5% of body weight over the 3 years.

"Even with the substantial differences between the participant and nonparticipant groups, direct comparisons of the risk factors are hard to interpret," Dr. Jackson noted. For this reason, she and her associates used least-square means to compute averages, and adjusted for differences in baseline CVD risk factors, as well as differences in baseline BMI, age, gender, and race/ethnicity between the populations. They also stratified the analyses by diabetes status.

Systolic BP was a little lower among MOVE! participants than among nonparticipants, particularly at 6 and 12 months, "although the effect disappeared over 24 and 36 months," Dr. Jackson said. HDL cholesterol was a little higher among participants, compared with nonparticipants, while non-HDL cholesterol and plasma glucose levels were a little lower among participants, compared with nonparticipants.

Cox proportional hazards modeling revealed that there was a 17% lower incidence of CVD among participants, compared with nonparticipants (hazard ratio, 0.83). "When we looked separately at coronary artery disease, stroke, peripheral vascular disease, and heart failure, the effects remained significant," she said. These models were also adjusted for age, gender, BMI, smoking status, and a propensity score indicating likelihood of MOVE! participation, as well as other demographic and clinical characteristics.

A subgroup analysis revealed that the effect of MOVE! participation was stronger for men than for women (HR 0.83 vs. HR 0.93, respectively; P = .001).

"The VA population does have substantially fewer women than men, and women in the VA tend to be younger and may be at lower CVD risk than men," Dr. Jackson noted. "We also saw a stronger effect of participation among VA patients without diabetes, compared with those with diabetes, and we saw a stronger effect of participation among current smokers compared with nonsmokers."

Dr. Jackson said that she had no relevant financial conflicts to disclose.

dbrunk@frontlinemedcom.com

On Twitter @dougbrunk