Michele G. Sullivan

By stressing the importance of a carbohydrate-based, low-fat diet, current U.S. dietary guidelines may have unexpectedly contributed to the current obesity epidemic, investigators reported.

In accordance with national recommendations, Americans have slightly reduced their fat intake, wrote Dr. Paul Marantz of the Albert Einstein College of Medicine, New York, and his coauthors. But their carbohydrate and total-calorie intakes have increased, along with the rate of national obesity (Am. J. Prev. Med. 2008 Feb. 8 [Epub doi:10.1016/j.amepre.2007.11.017]).

The observation is not enough to establish a causal link, but enough data exist to make at least an inference. “The hypothesis that dietary fat admonitions actually caused the current U.S. obesity epidemic is consistent with the data, logically sound, and plausible on the basis of both behavioral and biological mechanisms,” they said.

The recommendation to reduce fat intake, first promulgated in 1980, focused on the association between cardiovascular disease and one risk factor: hypercholesterolemia. But although there was solid evidence that modifying fat intake could reduce cholesterol, there was—and still is—no evidence that governmental guidelines against fat could improve cardiovascular disease outcomes, the investigators said.

Instead, Dr. Marantz and his team argue, data now suggest that these guidelines negatively affected health by contributing to the obesity epidemic and its attendant increase in diabetes. They used statistics from the Centers for Disease Control and Prevention to support that view.

From 1971 to 2001, consistent with national recommendations of a low-fat, carbohydrate-based diet, fat intake decreased by 5% in men and 9% in women. But carbohydrate intake increased by 7% in men and 6% in women, and total daily caloric intake increased by 168 calories in men and 335 calories in women. In fact, even though women decreased their percentage of fat intake, their increase in daily calories translated into an increase in absolute fat intake, from 557 fat calories per day to 616 fat calories per day.

A corresponding increase in obesity ensued in both genders, the authors noted. In 1971, 55% of American men and 41% of women were overweight or obese; by 2001, those numbers had risen to 70% of men and 62% of women.

The relationship between the guidelines and changing dietary habits is probably multifactorial, they said. Fat may induce satiety—an important inhibitor of excess calorie intake—which would be a biologically plausible rationale for the idea that low-fat diets may lead to higher calorie consumption.

A societal force is probably also at work, they said. Total calorie intake “may have been influenced by the effective marketing of low-fat foods, as well as the food pyramid, which suggested that low-fat foods could be eaten without any concern,” and gave an official “seal of approval” for such foods.

The United States has enjoyed a decrease in the rates of cardiovascular mortality since the national low-fat recommendation was first made, the authors noted. “Of course this decline had begun in the 1960s prior to the dietary guidelines, and other clinical interventions (statins, bypass surgery, and angioplasty) also contributed. Moreover, this favorable trend in coronary heart disease was counterbalanced by an alarming increase in obesity and attendant diabetes that coincided with the promulgation of the 1980 dietary guidelines.”

The authors maintained that dietary guidelines should include explicit standards of evidence, such as the standards employed by the U.S. Preventive Services Task Force. “This may lead to guidelines that are laden with caveats and disclaimers, but these are preferable to resolute guidelines supported by equivocal evidence,” the investigators said. “When the evidence is murky, public health officials may be served best by … making no recommendation at all.”

An accompanying editorial by Dr. Steven Woolf and Marion Nestle, Ph.D., strongly challenges these conclusions. Current guidelines are based on dozens of randomized controlled trials linking low-fat diets with decreased disease risk, said Dr. Woolf of Virginia Commonwealth University, Richmond, and Dr. Nestle of New York University, New York. The insinuation that dietary guidelines contributed to increasing obesity also fails to account for other significant factors, like portion sizes, inactivity, and overall caloric intake, they said (Am. J. Prev. Med. 2008 Feb. 8 [Epub doi:10.1016/j.amepre.2007.12.002]).

Despite the improvements in cardiovascular disease mortality, two diseases linked to fat intake—cardiovascular disease and cancer—still account for more than half of U.S. deaths. Obesity and diabetes are expected to increase. Withholding guidelines because of fear of unintended consequences is not the answer to the obesity problems, according to Dr. Woolf and Dr. Nestle.

“Under these circumstances, the public is placed at greater risk by withholding information about dietary causes than by sharing it. Withholding dietary guidance out of fear of unintended consequences elevates the duty for caution above the duty to inform,” they wrote in the editorial.

None of the authors of the article and the editorial reported any financial disclosures.

By stressing the importance of a carbohydrate-based, low-fat diet, current U.S. dietary guidelines may have unexpectedly contributed to the current obesity epidemic, investigators reported.

In accordance with national recommendations, Americans have slightly reduced their fat intake, wrote Dr. Paul Marantz of the Albert Einstein College of Medicine, New York, and his coauthors. But their carbohydrate and total-calorie intakes have increased, along with the rate of national obesity (Am. J. Prev. Med. 2008 Feb. 8 [Epub doi:10.1016/j.amepre.2007.11.017]).

The observation is not enough to establish a causal link, but enough data exist to make at least an inference. “The hypothesis that dietary fat admonitions actually caused the current U.S. obesity epidemic is consistent with the data, logically sound, and plausible on the basis of both behavioral and biological mechanisms,” they said.

The recommendation to reduce fat intake, first promulgated in 1980, focused on the association between cardiovascular disease and one risk factor: hypercholesterolemia. But although there was solid evidence that modifying fat intake could reduce cholesterol, there was—and still is—no evidence that governmental guidelines against fat could improve cardiovascular disease outcomes, the investigators said.

Instead, Dr. Marantz and his team argue, data now suggest that these guidelines negatively affected health by contributing to the obesity epidemic and its attendant increase in diabetes. They used statistics from the Centers for Disease Control and Prevention to support that view.

From 1971 to 2001, consistent with national recommendations of a low-fat, carbohydrate-based diet, fat intake decreased by 5% in men and 9% in women. But carbohydrate intake increased by 7% in men and 6% in women, and total daily caloric intake increased by 168 calories in men and 335 calories in women. In fact, even though women decreased their percentage of fat intake, their increase in daily calories translated into an increase in absolute fat intake, from 557 fat calories per day to 616 fat calories per day.

A corresponding increase in obesity ensued in both genders, the authors noted. In 1971, 55% of American men and 41% of women were overweight or obese; by 2001, those numbers had risen to 70% of men and 62% of women.

The relationship between the guidelines and changing dietary habits is probably multifactorial, they said. Fat may induce satiety—an important inhibitor of excess calorie intake—which would be a biologically plausible rationale for the idea that low-fat diets may lead to higher calorie consumption.

A societal force is probably also at work, they said. Total calorie intake “may have been influenced by the effective marketing of low-fat foods, as well as the food pyramid, which suggested that low-fat foods could be eaten without any concern,” and gave an official “seal of approval” for such foods.

The United States has enjoyed a decrease in the rates of cardiovascular mortality since the national low-fat recommendation was first made, the authors noted. “Of course this decline had begun in the 1960s prior to the dietary guidelines, and other clinical interventions (statins, bypass surgery, and angioplasty) also contributed. Moreover, this favorable trend in coronary heart disease was counterbalanced by an alarming increase in obesity and attendant diabetes that coincided with the promulgation of the 1980 dietary guidelines.”

The authors maintained that dietary guidelines should include explicit standards of evidence, such as the standards employed by the U.S. Preventive Services Task Force. “This may lead to guidelines that are laden with caveats and disclaimers, but these are preferable to resolute guidelines supported by equivocal evidence,” the investigators said. “When the evidence is murky, public health officials may be served best by … making no recommendation at all.”

An accompanying editorial by Dr. Steven Woolf and Marion Nestle, Ph.D., strongly challenges these conclusions. Current guidelines are based on dozens of randomized controlled trials linking low-fat diets with decreased disease risk, said Dr. Woolf of Virginia Commonwealth University, Richmond, and Dr. Nestle of New York University, New York. The insinuation that dietary guidelines contributed to increasing obesity also fails to account for other significant factors, like portion sizes, inactivity, and overall caloric intake, they said (Am. J. Prev. Med. 2008 Feb. 8 [Epub doi:10.1016/j.amepre.2007.12.002]).

Despite the improvements in cardiovascular disease mortality, two diseases linked to fat intake—cardiovascular disease and cancer—still account for more than half of U.S. deaths. Obesity and diabetes are expected to increase. Withholding guidelines because of fear of unintended consequences is not the answer to the obesity problems, according to Dr. Woolf and Dr. Nestle.

“Under these circumstances, the public is placed at greater risk by withholding information about dietary causes than by sharing it. Withholding dietary guidance out of fear of unintended consequences elevates the duty for caution above the duty to inform,” they wrote in the editorial.

None of the authors of the article and the editorial reported any financial disclosures.

By stressing the importance of a carbohydrate-based, low-fat diet, current U.S. dietary guidelines may have unexpectedly contributed to the current obesity epidemic, investigators reported.

In accordance with national recommendations, Americans have slightly reduced their fat intake, wrote Dr. Paul Marantz of the Albert Einstein College of Medicine, New York, and his coauthors. But their carbohydrate and total-calorie intakes have increased, along with the rate of national obesity (Am. J. Prev. Med. 2008 Feb. 8 [Epub doi:10.1016/j.amepre.2007.11.017]).

The observation is not enough to establish a causal link, but enough data exist to make at least an inference. “The hypothesis that dietary fat admonitions actually caused the current U.S. obesity epidemic is consistent with the data, logically sound, and plausible on the basis of both behavioral and biological mechanisms,” they said.

The recommendation to reduce fat intake, first promulgated in 1980, focused on the association between cardiovascular disease and one risk factor: hypercholesterolemia. But although there was solid evidence that modifying fat intake could reduce cholesterol, there was—and still is—no evidence that governmental guidelines against fat could improve cardiovascular disease outcomes, the investigators said.

Instead, Dr. Marantz and his team argue, data now suggest that these guidelines negatively affected health by contributing to the obesity epidemic and its attendant increase in diabetes. They used statistics from the Centers for Disease Control and Prevention to support that view.

From 1971 to 2001, consistent with national recommendations of a low-fat, carbohydrate-based diet, fat intake decreased by 5% in men and 9% in women. But carbohydrate intake increased by 7% in men and 6% in women, and total daily caloric intake increased by 168 calories in men and 335 calories in women. In fact, even though women decreased their percentage of fat intake, their increase in daily calories translated into an increase in absolute fat intake, from 557 fat calories per day to 616 fat calories per day.

A corresponding increase in obesity ensued in both genders, the authors noted. In 1971, 55% of American men and 41% of women were overweight or obese; by 2001, those numbers had risen to 70% of men and 62% of women.

The relationship between the guidelines and changing dietary habits is probably multifactorial, they said. Fat may induce satiety—an important inhibitor of excess calorie intake—which would be a biologically plausible rationale for the idea that low-fat diets may lead to higher calorie consumption.

A societal force is probably also at work, they said. Total calorie intake “may have been influenced by the effective marketing of low-fat foods, as well as the food pyramid, which suggested that low-fat foods could be eaten without any concern,” and gave an official “seal of approval” for such foods.

The United States has enjoyed a decrease in the rates of cardiovascular mortality since the national low-fat recommendation was first made, the authors noted. “Of course this decline had begun in the 1960s prior to the dietary guidelines, and other clinical interventions (statins, bypass surgery, and angioplasty) also contributed. Moreover, this favorable trend in coronary heart disease was counterbalanced by an alarming increase in obesity and attendant diabetes that coincided with the promulgation of the 1980 dietary guidelines.”

The authors maintained that dietary guidelines should include explicit standards of evidence, such as the standards employed by the U.S. Preventive Services Task Force. “This may lead to guidelines that are laden with caveats and disclaimers, but these are preferable to resolute guidelines supported by equivocal evidence,” the investigators said. “When the evidence is murky, public health officials may be served best by … making no recommendation at all.”

An accompanying editorial by Dr. Steven Woolf and Marion Nestle, Ph.D., strongly challenges these conclusions. Current guidelines are based on dozens of randomized controlled trials linking low-fat diets with decreased disease risk, said Dr. Woolf of Virginia Commonwealth University, Richmond, and Dr. Nestle of New York University, New York. The insinuation that dietary guidelines contributed to increasing obesity also fails to account for other significant factors, like portion sizes, inactivity, and overall caloric intake, they said (Am. J. Prev. Med. 2008 Feb. 8 [Epub doi:10.1016/j.amepre.2007.12.002]).

Despite the improvements in cardiovascular disease mortality, two diseases linked to fat intake—cardiovascular disease and cancer—still account for more than half of U.S. deaths. Obesity and diabetes are expected to increase. Withholding guidelines because of fear of unintended consequences is not the answer to the obesity problems, according to Dr. Woolf and Dr. Nestle.

“Under these circumstances, the public is placed at greater risk by withholding information about dietary causes than by sharing it. Withholding dietary guidance out of fear of unintended consequences elevates the duty for caution above the duty to inform,” they wrote in the editorial.

None of the authors of the article and the editorial reported any financial disclosures.

Women who have an abnormal glucose tolerance test result during pregnancy but do not develop gestational diabetes still face an increased risk of developing type 2 diabetes later on, investigators reported.

The large retrospective study, published Jan. 25, concluded that even modestly elevated glucose levels double the risk of diabetes within the next 9 years. “The risk of subsequent diabetes … likely occurs since these women have an intermediate form of glucose intolerance” with impaired β-cell functioning, wrote Dr. Darcy B. Carr of the University of Washington, Seattle, and her coauthors (Diabetes Care 2008 Jan. 25 [doi 10.2337/dc07-1957]).

In this retrospective cohort study, the researchers analyzed diabetes risk over a mean 9-year follow-up period in 31,000 women without gestational diabetes who had an oral glucose tolerance test (OGTT) or oral glucose challenge test (OGCT) during their pregnancy. The mean age was 31 years; the median follow-up was 9 years.

The investigators found that the risk of later development of type 2 diabetes rose as the values of the OGCT rose. Compared with women whose levels were normal, women with glucose levels of 5.4–6.2 mmol/L and 6.4–7.3 mmol/L had double the risk of developing the disease, while women with levels greater than 7.3 mmol/L were three times more likely to do so. Women with no abnormal values on the OGTT were at no increased risk of developing type 2 diabetes, but those with one abnormal value were twice as likely to do so.

These associations remained significant even after the researchers controlled for age, primigravidity, and preterm delivery.

The finding is consistent with those from a previous, much smaller longitudinal study that reported higher frequencies of glucose intolerance in women with one abnormal OGTT value.

Dr. Carr and her colleagues noted that their study could not control for race, family history, or body mass index—all important factors to consider when assessing diabetes risk. In addition, subsequent diabetes was not systematically assessed, which may introduce bias in those who were selected for testing, they wrote.

They also said their conclusions are not sufficient for them to make any screening or treatment recommendations, adding that, “Whether women who fall within this intermediate range of glucose intolerance during pregnancy may benefit from increased diabetes surveillance as well as lifestyle recommendations proven to reduce the risk of developing diabetes is unknown.”

Women who have an abnormal glucose tolerance test result during pregnancy but do not develop gestational diabetes still face an increased risk of developing type 2 diabetes later on, investigators reported.

The large retrospective study, published Jan. 25, concluded that even modestly elevated glucose levels double the risk of diabetes within the next 9 years. “The risk of subsequent diabetes … likely occurs since these women have an intermediate form of glucose intolerance” with impaired β-cell functioning, wrote Dr. Darcy B. Carr of the University of Washington, Seattle, and her coauthors (Diabetes Care 2008 Jan. 25 [doi 10.2337/dc07-1957]).

In this retrospective cohort study, the researchers analyzed diabetes risk over a mean 9-year follow-up period in 31,000 women without gestational diabetes who had an oral glucose tolerance test (OGTT) or oral glucose challenge test (OGCT) during their pregnancy. The mean age was 31 years; the median follow-up was 9 years.

The investigators found that the risk of later development of type 2 diabetes rose as the values of the OGCT rose. Compared with women whose levels were normal, women with glucose levels of 5.4–6.2 mmol/L and 6.4–7.3 mmol/L had double the risk of developing the disease, while women with levels greater than 7.3 mmol/L were three times more likely to do so. Women with no abnormal values on the OGTT were at no increased risk of developing type 2 diabetes, but those with one abnormal value were twice as likely to do so.

These associations remained significant even after the researchers controlled for age, primigravidity, and preterm delivery.

The finding is consistent with those from a previous, much smaller longitudinal study that reported higher frequencies of glucose intolerance in women with one abnormal OGTT value.

Dr. Carr and her colleagues noted that their study could not control for race, family history, or body mass index—all important factors to consider when assessing diabetes risk. In addition, subsequent diabetes was not systematically assessed, which may introduce bias in those who were selected for testing, they wrote.

They also said their conclusions are not sufficient for them to make any screening or treatment recommendations, adding that, “Whether women who fall within this intermediate range of glucose intolerance during pregnancy may benefit from increased diabetes surveillance as well as lifestyle recommendations proven to reduce the risk of developing diabetes is unknown.”

Women who have an abnormal glucose tolerance test result during pregnancy but do not develop gestational diabetes still face an increased risk of developing type 2 diabetes later on, investigators reported.

The large retrospective study, published Jan. 25, concluded that even modestly elevated glucose levels double the risk of diabetes within the next 9 years. “The risk of subsequent diabetes … likely occurs since these women have an intermediate form of glucose intolerance” with impaired β-cell functioning, wrote Dr. Darcy B. Carr of the University of Washington, Seattle, and her coauthors (Diabetes Care 2008 Jan. 25 [doi 10.2337/dc07-1957]).

In this retrospective cohort study, the researchers analyzed diabetes risk over a mean 9-year follow-up period in 31,000 women without gestational diabetes who had an oral glucose tolerance test (OGTT) or oral glucose challenge test (OGCT) during their pregnancy. The mean age was 31 years; the median follow-up was 9 years.

The investigators found that the risk of later development of type 2 diabetes rose as the values of the OGCT rose. Compared with women whose levels were normal, women with glucose levels of 5.4–6.2 mmol/L and 6.4–7.3 mmol/L had double the risk of developing the disease, while women with levels greater than 7.3 mmol/L were three times more likely to do so. Women with no abnormal values on the OGTT were at no increased risk of developing type 2 diabetes, but those with one abnormal value were twice as likely to do so.

These associations remained significant even after the researchers controlled for age, primigravidity, and preterm delivery.

The finding is consistent with those from a previous, much smaller longitudinal study that reported higher frequencies of glucose intolerance in women with one abnormal OGTT value.

Dr. Carr and her colleagues noted that their study could not control for race, family history, or body mass index—all important factors to consider when assessing diabetes risk. In addition, subsequent diabetes was not systematically assessed, which may introduce bias in those who were selected for testing, they wrote.

They also said their conclusions are not sufficient for them to make any screening or treatment recommendations, adding that, “Whether women who fall within this intermediate range of glucose intolerance during pregnancy may benefit from increased diabetes surveillance as well as lifestyle recommendations proven to reduce the risk of developing diabetes is unknown.”

Vascular abnormalities of the placenta were strongly correlated with the incidence of infantile hemangiomas in a small group of very-low-birth-weight infants, Dr. Juan Carlos Lopez Gutierrez and his colleagues reported.

Abnormalities such as infarcts and hematomas might create a hypoxic intraplacental environment that stimulates vasculogenesis and predisposes these infants to postnatal hemangioma growth, said Dr. Lopez and his associates of La Paz Children's Hospital, Madrid (Pediatr. Dermatol. 2007;24:353–5).

Although the study is too small to make absolute claims, it does seem to shed additional light on the prevailing theory of hemangioma pathogenesis: embolization of placental endothelial cells. “There are several different factors playing an important role in the pathogenesis of hemangiomas. Just embolization of placental cells as the origin of hemangiomas is too simple a theory,” Dr. Lopez said in an interview.

His case series suggests that placental lesions lead to a hypoxic environment that, in turn, stimulates the unopposed growth of escaped fetal endothelial cells. “Hypoxia is extremely important as a precursor lesion, and placental anomalies provoke a low-oxygen atmosphere, which is one—and probably the most important—of the suggested factors for hemangioma development,” Dr. Lopez said. His study included 26 very-low-birth-weight infants, 13 of whom developed infantile hemangiomas postnatally. The investigators examined each placenta macroscopically and microscopically.

Every placenta in the group of infants who developed hemangiomas was abnormal, they found. Two placentas showed massive retroplacental hematomas; seven showed extensive infarction; and four exhibited large dilated vascular cavitations, severe vasculitis, chorioamnionitis, and funiculitis. Cord insertion was marginal in eight, paracentral in three, and velamentous in two. Among the placentas showing infarcts, the area of ischemia was simple in two and multiple in five, with a mean infarct size of 15 mm. The infarcted areas resulted in decreased growth of peripheral villi in all tissues.

In contrast, all of the placentas of the infants without hemangiomas were free of lesions reflecting disturbed maternal-fetal circulation. Two placentas showed isolated villous immaturity. Cord insertion was central in four and paracentral in nine.

Hypoxia is an important angiogenic stimulator in fetal development. “The relatively low oxygen environment in which the human fetoplacental unit develops during the first trimester is necessary to induce vasculoangiogenesis via embryonic endothelial cells proliferation, as these cells are sensitive to hypoxia and acidosis,” Dr. Lopez and his associates noted.

Prior research suggests that abnormal placentas increase the likelihood of infantile hemangiomas through shearing and embolization of placental tissue. Thus, the hypoxic environment created by placental insufficiency could be the trigger that turns on a vasculogenic response in any endothelial cells that have escaped into a fetus's circulation, they wrote.

Dr. Paula North, chief of pediatric pathology at the Medical College of Wisconsin, Milwaukee, pioneered the embolization theory of hemangioma pathogenesis and said that Dr. Lopez's study raises some valid issues. “The evidence from this study is a little bit circumstantial in supporting the placental origin theory,” Dr. North said in an interview. “What it does suggest is the idea that any kind of placental injury would increase the shedding of vascular precursor cells from the placenta, which then migrate into the baby.”

Once the placental cells enter the fetus, their migratory path and growth are probably influenced by their phenotype. “The cells that make up a hemangioma express a very interesting pattern of molecules that are highly relevant to the immune system,” Dr. North said. “They express indoleamine 2,3-dioxygenase, which helps create a state of maternal immune tolerance to the fetus, and this could help protect the growing hemangioma from attack by activated T cells.”

Hemangioma cells also express chemokine receptor 6. Normally expressed in dendritic cells, it might influence the area where shed placental endothelial cells eventually lodge, working “like a homing mechanism to bring the cells to the skin and liver,” she said.

Vascular abnormalities of the placenta were strongly correlated with the incidence of infantile hemangiomas in a small group of very-low-birth-weight infants, Dr. Juan Carlos Lopez Gutierrez and his colleagues reported.

Abnormalities such as infarcts and hematomas might create a hypoxic intraplacental environment that stimulates vasculogenesis and predisposes these infants to postnatal hemangioma growth, said Dr. Lopez and his associates of La Paz Children's Hospital, Madrid (Pediatr. Dermatol. 2007;24:353–5).

Although the study is too small to make absolute claims, it does seem to shed additional light on the prevailing theory of hemangioma pathogenesis: embolization of placental endothelial cells. “There are several different factors playing an important role in the pathogenesis of hemangiomas. Just embolization of placental cells as the origin of hemangiomas is too simple a theory,” Dr. Lopez said in an interview.

His case series suggests that placental lesions lead to a hypoxic environment that, in turn, stimulates the unopposed growth of escaped fetal endothelial cells. “Hypoxia is extremely important as a precursor lesion, and placental anomalies provoke a low-oxygen atmosphere, which is one—and probably the most important—of the suggested factors for hemangioma development,” Dr. Lopez said. His study included 26 very-low-birth-weight infants, 13 of whom developed infantile hemangiomas postnatally. The investigators examined each placenta macroscopically and microscopically.

Every placenta in the group of infants who developed hemangiomas was abnormal, they found. Two placentas showed massive retroplacental hematomas; seven showed extensive infarction; and four exhibited large dilated vascular cavitations, severe vasculitis, chorioamnionitis, and funiculitis. Cord insertion was marginal in eight, paracentral in three, and velamentous in two. Among the placentas showing infarcts, the area of ischemia was simple in two and multiple in five, with a mean infarct size of 15 mm. The infarcted areas resulted in decreased growth of peripheral villi in all tissues.

In contrast, all of the placentas of the infants without hemangiomas were free of lesions reflecting disturbed maternal-fetal circulation. Two placentas showed isolated villous immaturity. Cord insertion was central in four and paracentral in nine.

Hypoxia is an important angiogenic stimulator in fetal development. “The relatively low oxygen environment in which the human fetoplacental unit develops during the first trimester is necessary to induce vasculoangiogenesis via embryonic endothelial cells proliferation, as these cells are sensitive to hypoxia and acidosis,” Dr. Lopez and his associates noted.

Prior research suggests that abnormal placentas increase the likelihood of infantile hemangiomas through shearing and embolization of placental tissue. Thus, the hypoxic environment created by placental insufficiency could be the trigger that turns on a vasculogenic response in any endothelial cells that have escaped into a fetus's circulation, they wrote.

Dr. Paula North, chief of pediatric pathology at the Medical College of Wisconsin, Milwaukee, pioneered the embolization theory of hemangioma pathogenesis and said that Dr. Lopez's study raises some valid issues. “The evidence from this study is a little bit circumstantial in supporting the placental origin theory,” Dr. North said in an interview. “What it does suggest is the idea that any kind of placental injury would increase the shedding of vascular precursor cells from the placenta, which then migrate into the baby.”

Once the placental cells enter the fetus, their migratory path and growth are probably influenced by their phenotype. “The cells that make up a hemangioma express a very interesting pattern of molecules that are highly relevant to the immune system,” Dr. North said. “They express indoleamine 2,3-dioxygenase, which helps create a state of maternal immune tolerance to the fetus, and this could help protect the growing hemangioma from attack by activated T cells.”

Hemangioma cells also express chemokine receptor 6. Normally expressed in dendritic cells, it might influence the area where shed placental endothelial cells eventually lodge, working “like a homing mechanism to bring the cells to the skin and liver,” she said.

Vascular abnormalities of the placenta were strongly correlated with the incidence of infantile hemangiomas in a small group of very-low-birth-weight infants, Dr. Juan Carlos Lopez Gutierrez and his colleagues reported.

Abnormalities such as infarcts and hematomas might create a hypoxic intraplacental environment that stimulates vasculogenesis and predisposes these infants to postnatal hemangioma growth, said Dr. Lopez and his associates of La Paz Children's Hospital, Madrid (Pediatr. Dermatol. 2007;24:353–5).

Although the study is too small to make absolute claims, it does seem to shed additional light on the prevailing theory of hemangioma pathogenesis: embolization of placental endothelial cells. “There are several different factors playing an important role in the pathogenesis of hemangiomas. Just embolization of placental cells as the origin of hemangiomas is too simple a theory,” Dr. Lopez said in an interview.

His case series suggests that placental lesions lead to a hypoxic environment that, in turn, stimulates the unopposed growth of escaped fetal endothelial cells. “Hypoxia is extremely important as a precursor lesion, and placental anomalies provoke a low-oxygen atmosphere, which is one—and probably the most important—of the suggested factors for hemangioma development,” Dr. Lopez said. His study included 26 very-low-birth-weight infants, 13 of whom developed infantile hemangiomas postnatally. The investigators examined each placenta macroscopically and microscopically.

Every placenta in the group of infants who developed hemangiomas was abnormal, they found. Two placentas showed massive retroplacental hematomas; seven showed extensive infarction; and four exhibited large dilated vascular cavitations, severe vasculitis, chorioamnionitis, and funiculitis. Cord insertion was marginal in eight, paracentral in three, and velamentous in two. Among the placentas showing infarcts, the area of ischemia was simple in two and multiple in five, with a mean infarct size of 15 mm. The infarcted areas resulted in decreased growth of peripheral villi in all tissues.

In contrast, all of the placentas of the infants without hemangiomas were free of lesions reflecting disturbed maternal-fetal circulation. Two placentas showed isolated villous immaturity. Cord insertion was central in four and paracentral in nine.

Hypoxia is an important angiogenic stimulator in fetal development. “The relatively low oxygen environment in which the human fetoplacental unit develops during the first trimester is necessary to induce vasculoangiogenesis via embryonic endothelial cells proliferation, as these cells are sensitive to hypoxia and acidosis,” Dr. Lopez and his associates noted.

Prior research suggests that abnormal placentas increase the likelihood of infantile hemangiomas through shearing and embolization of placental tissue. Thus, the hypoxic environment created by placental insufficiency could be the trigger that turns on a vasculogenic response in any endothelial cells that have escaped into a fetus's circulation, they wrote.

Dr. Paula North, chief of pediatric pathology at the Medical College of Wisconsin, Milwaukee, pioneered the embolization theory of hemangioma pathogenesis and said that Dr. Lopez's study raises some valid issues. “The evidence from this study is a little bit circumstantial in supporting the placental origin theory,” Dr. North said in an interview. “What it does suggest is the idea that any kind of placental injury would increase the shedding of vascular precursor cells from the placenta, which then migrate into the baby.”

Once the placental cells enter the fetus, their migratory path and growth are probably influenced by their phenotype. “The cells that make up a hemangioma express a very interesting pattern of molecules that are highly relevant to the immune system,” Dr. North said. “They express indoleamine 2,3-dioxygenase, which helps create a state of maternal immune tolerance to the fetus, and this could help protect the growing hemangioma from attack by activated T cells.”

Hemangioma cells also express chemokine receptor 6. Normally expressed in dendritic cells, it might influence the area where shed placental endothelial cells eventually lodge, working “like a homing mechanism to bring the cells to the skin and liver,” she said.

WASHINGTON — A new subanalysis of the TRITON-TIMI 38 trial shows that prasugrel significantly reduced the risk of death and subsequent heart attack in patients who had a first MI during the study.

Examining the full response of patients who have an event in any trial provides valuable information about a drug's total effect, Dr. Eugene Braunwald reported at a symposium sponsored by the Cardiovascular Research Institute at the Washington Hospital Center. Most patients who have a primary end point event will survive and stay in the study, although only that first event is included in the outcome analysis.

“The classical analysis we do in all clinical trials doesn't tell the whole story—the number of total events. A subanalysis of what happens to these patients tells you a lot about what the drug actually does,” said Dr. Braunwald of Brigham and Women's Hospital, Boston, and chairman of the TIMI study group.

The TRITON-TIMI 38 (Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition With Prasugrel—Thrombolysis in Myocardial Infarction) randomized 13,608 patients who had moderate- to high-risk acute coronary syndromes and were scheduled for percutaneous coronary intervention to either the investigational thienopyridine prasugrel (60-mg loading dose followed by 10-mg daily maintenance dosage) or clopidogrel (300-mg loading dose followed by 75-mg daily maintenance dosage), for up to 15 months.

Those taking prasugrel were 19% less likely to experience cardiovascular death, MI, or stroke; the number needed to treat was 46. However, Dr. Braunwald noted, “there was no free lunch.” Prasugrel carried a 6% increased risk of major bleeding, with 35 more incidents than occurred in the clopidogrel group.

The new analysis concluded that even among those 1,424 patients who had a first primary end point event, prasugrel was significantly more effective than clopidogrel at reducing the risk of death and subsequent events.

“After the first event, patients who were assigned to clopidogrel had an additional 113 events, compared to an additional 56 events among those taking prasugrel,” Dr. Braunwald said. “That was a highly significant difference. Additionally, there were 40 deaths after the initial event in the clopidogrel group, but only 14 in the prasugrel group, for a significant risk reduction of 34%.”

He also presented analyses examining prasugrel's sustained reduction in the risk of secondary end point events in the entire study group. Over the 15-month study period, prasugrel reduced urgent target vessel revascularization by 34%. This significant difference appeared within the first 30 days (47% reduction) and continued throughout the next 14 months (21% reduction).

Overall, prasugrel reduced stent thrombosis by 52%, compared with clopidogrel. That difference also emerged quite early, in the first 3 days of administration (51% reduction).

The TIMI study group continues to analyze prasugrel's effect in patients with either bare metal or drug-eluting stents, Dr. Braunwald said. “The analyses are not complete, but I think we can say that the 50% reduction in stent thrombosis will be similar for both bare metal and drug-eluting tents.”

Dr. Braunwald stressed that excess bleeding is the price for more effective platelet inhibition. “While there were no significant differences in the number of instrumented bleeds, there were significantly more spontaneous bleeds. Although the absolute number of events was small, the difference was significant.”

Even the excess bleeding, however, doesn't significantly negate prasugrel's overall benefit, Dr. Deepak Bhatt said while commenting on Dr. Braunwald's presentation. “The net clinical benefit was very significantly in favor of prasugrel even when you consider the bleeding episodes,” said Dr. Bhatt of the Cleveland Clinic. Despite this, he said, it will be important to continue analyzing the incidences of bleeding in both groups.

“Other trials suggest a relationship between bleeding and mortality, apart from the obvious [death resulting from exsanguination]. Bleeding can lead to hypotension, transfusion that may or not be appropriate, and cessation of aspirin or any other form of antithrombotic therapy, either by the doctor or by the patient without the doctor's knowledge. By other pathways, bleeding may lead to ischemia, stent thrombosis, or inflammation, which may contribute to mortality.”

The trial's finding of an attenuated bleeding risk among patients with diabetes (about 3,000) is an intriguing one, Dr. Bhatt noted. In this group, the rates of bleeding were not different between the two study drugs (2.6% prasugrel vs. 2.5% clopidogrel). “This could be a chance finding due to the small patient group, but an alternative explanation is that patients like diabetics, whose platelets are already 'revved up,' are most likely to derive benefit from prasugrel and least likely to suffer harm from it.

Both Dr. Braunwald and Dr. Bhatt reported receiving grants from Daiichi Sankyo Co. and Eli Lilly & Co., sponsors of the TRITON-TIMI 38 study.

WASHINGTON — A new subanalysis of the TRITON-TIMI 38 trial shows that prasugrel significantly reduced the risk of death and subsequent heart attack in patients who had a first MI during the study.

Examining the full response of patients who have an event in any trial provides valuable information about a drug's total effect, Dr. Eugene Braunwald reported at a symposium sponsored by the Cardiovascular Research Institute at the Washington Hospital Center. Most patients who have a primary end point event will survive and stay in the study, although only that first event is included in the outcome analysis.

“The classical analysis we do in all clinical trials doesn't tell the whole story—the number of total events. A subanalysis of what happens to these patients tells you a lot about what the drug actually does,” said Dr. Braunwald of Brigham and Women's Hospital, Boston, and chairman of the TIMI study group.

The TRITON-TIMI 38 (Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition With Prasugrel—Thrombolysis in Myocardial Infarction) randomized 13,608 patients who had moderate- to high-risk acute coronary syndromes and were scheduled for percutaneous coronary intervention to either the investigational thienopyridine prasugrel (60-mg loading dose followed by 10-mg daily maintenance dosage) or clopidogrel (300-mg loading dose followed by 75-mg daily maintenance dosage), for up to 15 months.

Those taking prasugrel were 19% less likely to experience cardiovascular death, MI, or stroke; the number needed to treat was 46. However, Dr. Braunwald noted, “there was no free lunch.” Prasugrel carried a 6% increased risk of major bleeding, with 35 more incidents than occurred in the clopidogrel group.

The new analysis concluded that even among those 1,424 patients who had a first primary end point event, prasugrel was significantly more effective than clopidogrel at reducing the risk of death and subsequent events.

“After the first event, patients who were assigned to clopidogrel had an additional 113 events, compared to an additional 56 events among those taking prasugrel,” Dr. Braunwald said. “That was a highly significant difference. Additionally, there were 40 deaths after the initial event in the clopidogrel group, but only 14 in the prasugrel group, for a significant risk reduction of 34%.”

He also presented analyses examining prasugrel's sustained reduction in the risk of secondary end point events in the entire study group. Over the 15-month study period, prasugrel reduced urgent target vessel revascularization by 34%. This significant difference appeared within the first 30 days (47% reduction) and continued throughout the next 14 months (21% reduction).

Overall, prasugrel reduced stent thrombosis by 52%, compared with clopidogrel. That difference also emerged quite early, in the first 3 days of administration (51% reduction).

The TIMI study group continues to analyze prasugrel's effect in patients with either bare metal or drug-eluting stents, Dr. Braunwald said. “The analyses are not complete, but I think we can say that the 50% reduction in stent thrombosis will be similar for both bare metal and drug-eluting tents.”

Dr. Braunwald stressed that excess bleeding is the price for more effective platelet inhibition. “While there were no significant differences in the number of instrumented bleeds, there were significantly more spontaneous bleeds. Although the absolute number of events was small, the difference was significant.”

Even the excess bleeding, however, doesn't significantly negate prasugrel's overall benefit, Dr. Deepak Bhatt said while commenting on Dr. Braunwald's presentation. “The net clinical benefit was very significantly in favor of prasugrel even when you consider the bleeding episodes,” said Dr. Bhatt of the Cleveland Clinic. Despite this, he said, it will be important to continue analyzing the incidences of bleeding in both groups.

“Other trials suggest a relationship between bleeding and mortality, apart from the obvious [death resulting from exsanguination]. Bleeding can lead to hypotension, transfusion that may or not be appropriate, and cessation of aspirin or any other form of antithrombotic therapy, either by the doctor or by the patient without the doctor's knowledge. By other pathways, bleeding may lead to ischemia, stent thrombosis, or inflammation, which may contribute to mortality.”

The trial's finding of an attenuated bleeding risk among patients with diabetes (about 3,000) is an intriguing one, Dr. Bhatt noted. In this group, the rates of bleeding were not different between the two study drugs (2.6% prasugrel vs. 2.5% clopidogrel). “This could be a chance finding due to the small patient group, but an alternative explanation is that patients like diabetics, whose platelets are already 'revved up,' are most likely to derive benefit from prasugrel and least likely to suffer harm from it.

Both Dr. Braunwald and Dr. Bhatt reported receiving grants from Daiichi Sankyo Co. and Eli Lilly & Co., sponsors of the TRITON-TIMI 38 study.

WASHINGTON — A new subanalysis of the TRITON-TIMI 38 trial shows that prasugrel significantly reduced the risk of death and subsequent heart attack in patients who had a first MI during the study.

Examining the full response of patients who have an event in any trial provides valuable information about a drug's total effect, Dr. Eugene Braunwald reported at a symposium sponsored by the Cardiovascular Research Institute at the Washington Hospital Center. Most patients who have a primary end point event will survive and stay in the study, although only that first event is included in the outcome analysis.

“The classical analysis we do in all clinical trials doesn't tell the whole story—the number of total events. A subanalysis of what happens to these patients tells you a lot about what the drug actually does,” said Dr. Braunwald of Brigham and Women's Hospital, Boston, and chairman of the TIMI study group.

The TRITON-TIMI 38 (Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition With Prasugrel—Thrombolysis in Myocardial Infarction) randomized 13,608 patients who had moderate- to high-risk acute coronary syndromes and were scheduled for percutaneous coronary intervention to either the investigational thienopyridine prasugrel (60-mg loading dose followed by 10-mg daily maintenance dosage) or clopidogrel (300-mg loading dose followed by 75-mg daily maintenance dosage), for up to 15 months.

Those taking prasugrel were 19% less likely to experience cardiovascular death, MI, or stroke; the number needed to treat was 46. However, Dr. Braunwald noted, “there was no free lunch.” Prasugrel carried a 6% increased risk of major bleeding, with 35 more incidents than occurred in the clopidogrel group.

The new analysis concluded that even among those 1,424 patients who had a first primary end point event, prasugrel was significantly more effective than clopidogrel at reducing the risk of death and subsequent events.

“After the first event, patients who were assigned to clopidogrel had an additional 113 events, compared to an additional 56 events among those taking prasugrel,” Dr. Braunwald said. “That was a highly significant difference. Additionally, there were 40 deaths after the initial event in the clopidogrel group, but only 14 in the prasugrel group, for a significant risk reduction of 34%.”

He also presented analyses examining prasugrel's sustained reduction in the risk of secondary end point events in the entire study group. Over the 15-month study period, prasugrel reduced urgent target vessel revascularization by 34%. This significant difference appeared within the first 30 days (47% reduction) and continued throughout the next 14 months (21% reduction).

Overall, prasugrel reduced stent thrombosis by 52%, compared with clopidogrel. That difference also emerged quite early, in the first 3 days of administration (51% reduction).

The TIMI study group continues to analyze prasugrel's effect in patients with either bare metal or drug-eluting stents, Dr. Braunwald said. “The analyses are not complete, but I think we can say that the 50% reduction in stent thrombosis will be similar for both bare metal and drug-eluting tents.”

Dr. Braunwald stressed that excess bleeding is the price for more effective platelet inhibition. “While there were no significant differences in the number of instrumented bleeds, there were significantly more spontaneous bleeds. Although the absolute number of events was small, the difference was significant.”

Even the excess bleeding, however, doesn't significantly negate prasugrel's overall benefit, Dr. Deepak Bhatt said while commenting on Dr. Braunwald's presentation. “The net clinical benefit was very significantly in favor of prasugrel even when you consider the bleeding episodes,” said Dr. Bhatt of the Cleveland Clinic. Despite this, he said, it will be important to continue analyzing the incidences of bleeding in both groups.

“Other trials suggest a relationship between bleeding and mortality, apart from the obvious [death resulting from exsanguination]. Bleeding can lead to hypotension, transfusion that may or not be appropriate, and cessation of aspirin or any other form of antithrombotic therapy, either by the doctor or by the patient without the doctor's knowledge. By other pathways, bleeding may lead to ischemia, stent thrombosis, or inflammation, which may contribute to mortality.”

The trial's finding of an attenuated bleeding risk among patients with diabetes (about 3,000) is an intriguing one, Dr. Bhatt noted. In this group, the rates of bleeding were not different between the two study drugs (2.6% prasugrel vs. 2.5% clopidogrel). “This could be a chance finding due to the small patient group, but an alternative explanation is that patients like diabetics, whose platelets are already 'revved up,' are most likely to derive benefit from prasugrel and least likely to suffer harm from it.

Both Dr. Braunwald and Dr. Bhatt reported receiving grants from Daiichi Sankyo Co. and Eli Lilly & Co., sponsors of the TRITON-TIMI 38 study.

You might not expect a hospital to advertise its errors, but that's what the public wants. And that's what the New York City Health and Hospitals Corporation is doing, according to Alan Aviles, the group's president.

In July, Mr. Aviles announced that the group's 11 hospitals would publicize their overall mortality rates, heart attack mortality rates, and rates of nosocomial infections, including central line, ventilator-associated, and surgical site infections.

Nineteen states, including New York, have legislation requiring the public reporting of nosocomial rates. And legislation adopted in 2005 requires New York hospitals to report their incidence of central line bloodstream infections, and coronary artery bypass graft and colon surgery site infections, to the state health department.

It's unclear when that information might be made public and whether it will appear as aggregate or facility-specific information. But Mr. Aviles has taken the bull by the horns because hospitals can't do a better job until they can see the job they're already doing, he said in an interview.

“One of the biggest problems in this industry is the extent to which we keep this kind of quality-related data close to the vest. The practical result of that attitude is that we expect our physicians to make improvements while they're groping around in the dark. They never have the benefit of knowing what we—or others—are achieving and where we stand on that spectrum.”

Publicly disclosing what has always been considered a hospital's deepest secrets is the only way to fix them, he said.

Mr. Aviles has had his share of naysayers, including those within his own system who worried that public scrutiny could nick their competitive edge among the city's 60 hospitals. “There was concern that we could be impacted competitively if the public either misinterpreted the data or if the numbers aren't as good as those of the competition,” he said. “But people know that medical errors and hospital-acquired infections cause thousands of needless deaths each year. They know there needs to be significant focus and improvement on these, and this transparency can only help.”

To that end, Mr. Aviles and his team have set a lofty goal: By 2010, they want to have the safest hospitals in the country.

There's no mistaking the single-mindedness behind that goal, said Jim Conway, senior vice president of the Institute for Healthcare Improvement (IHI). “It's extraordinarily courageous and extraordinarily hard,” he said in an interview. “They're willing to be held accountable not only to their own staff, but to consumers, patients, and families.”

The nonprofit IHI supports transformational change in health care quality and safety, both in the United States and internationally. Its “Five Million Lives Campaign,” launched in 2006, aims to protect patients from 5 million incidents of medical harm by the end of 2008. To achieve that, at least 4,000 hospitals will have to commit to improving patient safety. At present, 3,500 are involved.

The lofty goals set by the IHI and Mr. Aviles' group are a hallmark of successful change in health care systems, Mr. Conway said. Another example is Ascension Health System, which comprises 65 hospitals across the United States and aims to eliminate all preventable harm in all of their hospitals by 2008. “We're seeing the fruits of that goal. In almost every tracked indicator, their performance is much better than almost any other system in the country. For pressure ulcers, for example, the rate in their lowest-performing hospital is one-sixth that of the national average,” Mr. Conway added.

Cincinnati Children's Hospital is trying to eliminate 80% of preventable serious harm, including hospital-acquired infections, by July 2008, according to Mr. Conway. Beth Israel Deaconess Medical Center in Boston has become the first hospital to post its 2007 Joint Commission Accreditation Survey findings on its Web site. The center also publicly posts its commitments to quality improvement, including the complete elimination of ventilator-associated pneumonia and central line infections.

“What drives organizations to accomplish goals like these?” Mr. Conway asked. “They have a great vision, and they have a solid sense of their current reality. They understand the gap between where they are and where they need to be, and they use that tension to drive change.”

Goals in these organizations are specific and measurable—eliminating 100% of central line infections by a certain date, for instance, as opposed to a broader aim of delivering the best health care.

“We've come to understand that 'some' is not a number and 'soon' is not a time,” Mr. Conway said. Just as importantly, everyone from the chief surgeon to housekeeping is considered responsible, he added.

Such efforts may be unusual now, but they are the wave of the future, Mr. Conway said. The public wants the information, states are requiring its disclosure, and the federal government is now refusing to pay for illnesses that could have been prevented—including hospital-acquired infections.

“We're going to see more and more hospitals taking responsibility like this,” he predicted. “This is the end of the beginning, and the beginning of something new.”

You might not expect a hospital to advertise its errors, but that's what the public wants. And that's what the New York City Health and Hospitals Corporation is doing, according to Alan Aviles, the group's president.

In July, Mr. Aviles announced that the group's 11 hospitals would publicize their overall mortality rates, heart attack mortality rates, and rates of nosocomial infections, including central line, ventilator-associated, and surgical site infections.

Nineteen states, including New York, have legislation requiring the public reporting of nosocomial rates. And legislation adopted in 2005 requires New York hospitals to report their incidence of central line bloodstream infections, and coronary artery bypass graft and colon surgery site infections, to the state health department.

It's unclear when that information might be made public and whether it will appear as aggregate or facility-specific information. But Mr. Aviles has taken the bull by the horns because hospitals can't do a better job until they can see the job they're already doing, he said in an interview.

“One of the biggest problems in this industry is the extent to which we keep this kind of quality-related data close to the vest. The practical result of that attitude is that we expect our physicians to make improvements while they're groping around in the dark. They never have the benefit of knowing what we—or others—are achieving and where we stand on that spectrum.”

Publicly disclosing what has always been considered a hospital's deepest secrets is the only way to fix them, he said.

Mr. Aviles has had his share of naysayers, including those within his own system who worried that public scrutiny could nick their competitive edge among the city's 60 hospitals. “There was concern that we could be impacted competitively if the public either misinterpreted the data or if the numbers aren't as good as those of the competition,” he said. “But people know that medical errors and hospital-acquired infections cause thousands of needless deaths each year. They know there needs to be significant focus and improvement on these, and this transparency can only help.”

To that end, Mr. Aviles and his team have set a lofty goal: By 2010, they want to have the safest hospitals in the country.

There's no mistaking the single-mindedness behind that goal, said Jim Conway, senior vice president of the Institute for Healthcare Improvement (IHI). “It's extraordinarily courageous and extraordinarily hard,” he said in an interview. “They're willing to be held accountable not only to their own staff, but to consumers, patients, and families.”

The nonprofit IHI supports transformational change in health care quality and safety, both in the United States and internationally. Its “Five Million Lives Campaign,” launched in 2006, aims to protect patients from 5 million incidents of medical harm by the end of 2008. To achieve that, at least 4,000 hospitals will have to commit to improving patient safety. At present, 3,500 are involved.

The lofty goals set by the IHI and Mr. Aviles' group are a hallmark of successful change in health care systems, Mr. Conway said. Another example is Ascension Health System, which comprises 65 hospitals across the United States and aims to eliminate all preventable harm in all of their hospitals by 2008. “We're seeing the fruits of that goal. In almost every tracked indicator, their performance is much better than almost any other system in the country. For pressure ulcers, for example, the rate in their lowest-performing hospital is one-sixth that of the national average,” Mr. Conway added.

Cincinnati Children's Hospital is trying to eliminate 80% of preventable serious harm, including hospital-acquired infections, by July 2008, according to Mr. Conway. Beth Israel Deaconess Medical Center in Boston has become the first hospital to post its 2007 Joint Commission Accreditation Survey findings on its Web site. The center also publicly posts its commitments to quality improvement, including the complete elimination of ventilator-associated pneumonia and central line infections.

“What drives organizations to accomplish goals like these?” Mr. Conway asked. “They have a great vision, and they have a solid sense of their current reality. They understand the gap between where they are and where they need to be, and they use that tension to drive change.”

Goals in these organizations are specific and measurable—eliminating 100% of central line infections by a certain date, for instance, as opposed to a broader aim of delivering the best health care.

“We've come to understand that 'some' is not a number and 'soon' is not a time,” Mr. Conway said. Just as importantly, everyone from the chief surgeon to housekeeping is considered responsible, he added.

Such efforts may be unusual now, but they are the wave of the future, Mr. Conway said. The public wants the information, states are requiring its disclosure, and the federal government is now refusing to pay for illnesses that could have been prevented—including hospital-acquired infections.

“We're going to see more and more hospitals taking responsibility like this,” he predicted. “This is the end of the beginning, and the beginning of something new.”

You might not expect a hospital to advertise its errors, but that's what the public wants. And that's what the New York City Health and Hospitals Corporation is doing, according to Alan Aviles, the group's president.

In July, Mr. Aviles announced that the group's 11 hospitals would publicize their overall mortality rates, heart attack mortality rates, and rates of nosocomial infections, including central line, ventilator-associated, and surgical site infections.

Nineteen states, including New York, have legislation requiring the public reporting of nosocomial rates. And legislation adopted in 2005 requires New York hospitals to report their incidence of central line bloodstream infections, and coronary artery bypass graft and colon surgery site infections, to the state health department.

It's unclear when that information might be made public and whether it will appear as aggregate or facility-specific information. But Mr. Aviles has taken the bull by the horns because hospitals can't do a better job until they can see the job they're already doing, he said in an interview.

“One of the biggest problems in this industry is the extent to which we keep this kind of quality-related data close to the vest. The practical result of that attitude is that we expect our physicians to make improvements while they're groping around in the dark. They never have the benefit of knowing what we—or others—are achieving and where we stand on that spectrum.”

Publicly disclosing what has always been considered a hospital's deepest secrets is the only way to fix them, he said.

Mr. Aviles has had his share of naysayers, including those within his own system who worried that public scrutiny could nick their competitive edge among the city's 60 hospitals. “There was concern that we could be impacted competitively if the public either misinterpreted the data or if the numbers aren't as good as those of the competition,” he said. “But people know that medical errors and hospital-acquired infections cause thousands of needless deaths each year. They know there needs to be significant focus and improvement on these, and this transparency can only help.”

To that end, Mr. Aviles and his team have set a lofty goal: By 2010, they want to have the safest hospitals in the country.

There's no mistaking the single-mindedness behind that goal, said Jim Conway, senior vice president of the Institute for Healthcare Improvement (IHI). “It's extraordinarily courageous and extraordinarily hard,” he said in an interview. “They're willing to be held accountable not only to their own staff, but to consumers, patients, and families.”

The nonprofit IHI supports transformational change in health care quality and safety, both in the United States and internationally. Its “Five Million Lives Campaign,” launched in 2006, aims to protect patients from 5 million incidents of medical harm by the end of 2008. To achieve that, at least 4,000 hospitals will have to commit to improving patient safety. At present, 3,500 are involved.

The lofty goals set by the IHI and Mr. Aviles' group are a hallmark of successful change in health care systems, Mr. Conway said. Another example is Ascension Health System, which comprises 65 hospitals across the United States and aims to eliminate all preventable harm in all of their hospitals by 2008. “We're seeing the fruits of that goal. In almost every tracked indicator, their performance is much better than almost any other system in the country. For pressure ulcers, for example, the rate in their lowest-performing hospital is one-sixth that of the national average,” Mr. Conway added.

Cincinnati Children's Hospital is trying to eliminate 80% of preventable serious harm, including hospital-acquired infections, by July 2008, according to Mr. Conway. Beth Israel Deaconess Medical Center in Boston has become the first hospital to post its 2007 Joint Commission Accreditation Survey findings on its Web site. The center also publicly posts its commitments to quality improvement, including the complete elimination of ventilator-associated pneumonia and central line infections.

“What drives organizations to accomplish goals like these?” Mr. Conway asked. “They have a great vision, and they have a solid sense of their current reality. They understand the gap between where they are and where they need to be, and they use that tension to drive change.”

Goals in these organizations are specific and measurable—eliminating 100% of central line infections by a certain date, for instance, as opposed to a broader aim of delivering the best health care.

“We've come to understand that 'some' is not a number and 'soon' is not a time,” Mr. Conway said. Just as importantly, everyone from the chief surgeon to housekeeping is considered responsible, he added.

Such efforts may be unusual now, but they are the wave of the future, Mr. Conway said. The public wants the information, states are requiring its disclosure, and the federal government is now refusing to pay for illnesses that could have been prevented—including hospital-acquired infections.

“We're going to see more and more hospitals taking responsibility like this,” he predicted. “This is the end of the beginning, and the beginning of something new.”

Women who have an abnormal glucose tolerance test result during pregnancy but do not develop gestational diabetes still face an increased risk of developing type 2 diabetes later on.

The retrospective study showed that even modestly elevated glucose levels double the risk of diabetes within the next 9 years. “The risk of subsequent diabetes … likely occurs since [they] have an intermediate form of glucose intolerance” with impaired β-cell functioning, said Dr. Darcy B. Carr of the University of Washington, Seattle, and coauthors (Diabetes Care 2008 Jan. 25 [doi 10.2337/dc07–1957

In this retrospective cohort study, the researchers analyzed diabetes risk over a mean 9-year follow-up period in 31,000 women without gestational diabetes who had an oral glucose tolerance test (OGTT) or oral glucose challenge test (OGCT) during their pregnancy. The mean age was 31 years; the median follow-up was 9 years.

They found that the risk of later development of type 2 diabetes rose as the OGCT values rose. Compared with women whose levels were normal, those with glucose levels of 5.4–6.2 mmol/L and 6.4–7.3 mmol/L had double the risk of developing the disease, while those with levels greater than 7.3 mmol/L were three times more likely to do so. Women with no abnormal OGTT values were at no increased risk of developing type 2 diabetes, but those with one abnormal value were twice as likely to do so. These associations remained significant even after controlling for age, primigravidity, and preterm delivery.

The finding is consistent with those from a previous, much smaller longitudinal study that reported higher frequencies of glucose intolerance in women with one abnormal OGTT value.

The authors noted that the study could not control for race, family history, or body mass index—all important factors in assessing diabetes risk. In addition, subsequent diabetes was not systematically assessed, which may introduce bias in those who were selected for testing, they wrote.

They also said their conclusions are not sufficient for them to make any screening or treatment recommendations, adding that, “Whether women who fall within this intermediate range of glucose intolerance during pregnancy may benefit from increased diabetes surveillance as well as lifestyle recommendations proven to reduce the risk of developing diabetes is unknown.”

ELSEVIER GLOBAL MEDICAL NEWS

Women who have an abnormal glucose tolerance test result during pregnancy but do not develop gestational diabetes still face an increased risk of developing type 2 diabetes later on.

The retrospective study showed that even modestly elevated glucose levels double the risk of diabetes within the next 9 years. “The risk of subsequent diabetes … likely occurs since [they] have an intermediate form of glucose intolerance” with impaired β-cell functioning, said Dr. Darcy B. Carr of the University of Washington, Seattle, and coauthors (Diabetes Care 2008 Jan. 25 [doi 10.2337/dc07–1957

In this retrospective cohort study, the researchers analyzed diabetes risk over a mean 9-year follow-up period in 31,000 women without gestational diabetes who had an oral glucose tolerance test (OGTT) or oral glucose challenge test (OGCT) during their pregnancy. The mean age was 31 years; the median follow-up was 9 years.

They found that the risk of later development of type 2 diabetes rose as the OGCT values rose. Compared with women whose levels were normal, those with glucose levels of 5.4–6.2 mmol/L and 6.4–7.3 mmol/L had double the risk of developing the disease, while those with levels greater than 7.3 mmol/L were three times more likely to do so. Women with no abnormal OGTT values were at no increased risk of developing type 2 diabetes, but those with one abnormal value were twice as likely to do so. These associations remained significant even after controlling for age, primigravidity, and preterm delivery.

The finding is consistent with those from a previous, much smaller longitudinal study that reported higher frequencies of glucose intolerance in women with one abnormal OGTT value.

The authors noted that the study could not control for race, family history, or body mass index—all important factors in assessing diabetes risk. In addition, subsequent diabetes was not systematically assessed, which may introduce bias in those who were selected for testing, they wrote.

They also said their conclusions are not sufficient for them to make any screening or treatment recommendations, adding that, “Whether women who fall within this intermediate range of glucose intolerance during pregnancy may benefit from increased diabetes surveillance as well as lifestyle recommendations proven to reduce the risk of developing diabetes is unknown.”

ELSEVIER GLOBAL MEDICAL NEWS

Women who have an abnormal glucose tolerance test result during pregnancy but do not develop gestational diabetes still face an increased risk of developing type 2 diabetes later on.

The retrospective study showed that even modestly elevated glucose levels double the risk of diabetes within the next 9 years. “The risk of subsequent diabetes … likely occurs since [they] have an intermediate form of glucose intolerance” with impaired β-cell functioning, said Dr. Darcy B. Carr of the University of Washington, Seattle, and coauthors (Diabetes Care 2008 Jan. 25 [doi 10.2337/dc07–1957

In this retrospective cohort study, the researchers analyzed diabetes risk over a mean 9-year follow-up period in 31,000 women without gestational diabetes who had an oral glucose tolerance test (OGTT) or oral glucose challenge test (OGCT) during their pregnancy. The mean age was 31 years; the median follow-up was 9 years.

They found that the risk of later development of type 2 diabetes rose as the OGCT values rose. Compared with women whose levels were normal, those with glucose levels of 5.4–6.2 mmol/L and 6.4–7.3 mmol/L had double the risk of developing the disease, while those with levels greater than 7.3 mmol/L were three times more likely to do so. Women with no abnormal OGTT values were at no increased risk of developing type 2 diabetes, but those with one abnormal value were twice as likely to do so. These associations remained significant even after controlling for age, primigravidity, and preterm delivery.

The finding is consistent with those from a previous, much smaller longitudinal study that reported higher frequencies of glucose intolerance in women with one abnormal OGTT value.

The authors noted that the study could not control for race, family history, or body mass index—all important factors in assessing diabetes risk. In addition, subsequent diabetes was not systematically assessed, which may introduce bias in those who were selected for testing, they wrote.

They also said their conclusions are not sufficient for them to make any screening or treatment recommendations, adding that, “Whether women who fall within this intermediate range of glucose intolerance during pregnancy may benefit from increased diabetes surveillance as well as lifestyle recommendations proven to reduce the risk of developing diabetes is unknown.”

ELSEVIER GLOBAL MEDICAL NEWS

Age 60 seems to be the defining year for many homozygous carriers of the apolipoprotein ϵ4 gene –the time when age-related changes in cognition focus more on memory and begin a steeper decline into mild cognitive impairment and, eventually Alzheimer's disease, according to new unpublished observations from a longitudinal study of apo ϵ4 carriers and normal controls.

“We saw normal age-related patterns of memory loss that occurred before age 60, but during this period, we didn't see any significant cognitive differences between the apo ϵ4 homozygotes, heterozygotes, and noncarriers,” said Dr. Richard Caselli, a lead investigator for the Arizona Apoϵ4 Cohort Longitudinal Study of Cognitively Normal Individuals. “But our latest information shows that at around age 60, a separation begins and continues for as long as we have been able to follow our subjects. There is a particular pattern of decline in apo ϵ4 homozygotes that tends to precede any diagnosis of mild cognitive impairment [MCI] or anything that can be seen with routine clinical brain imaging,” he said in an interview.

This pattern suggests that pathologic changes could be occurring earlier in homozygous ϵ4 carriers, although the exact nature of these changes, and their triggers, remain speculative, said Dr. Caselli, chairman of the department of neurology at the Mayo Clinic, Scottsdale, Ariz., and a member of the Arizona Alzheimer's Disease Consortium.

The Arizona cohort was initiated in 1994, and now includes more than 600 people, enrolled at ages 20–90 years, who have at least one first-degree relative with Alzheimer's disease. The subjects are all genotyped for the apo ϵ4 allele, and undergo extensive neuropsychological testing every 2 years.

Dr. Caselli's recent substudy focuses on 214 of these subjects aged 50–69 years. Almost half are apo ϵ4 carriers–43 homozygous and 59 heterozygous.

The study set out to characterize the effect of apo ϵ4 status on the development of presymptomatic cognitive changes. It's well known that the gene has a dose-response effect on the age at AD diagnosis: 80%–90% of homozygotes will develop the disease, at a mean age of 68 years. About 30% of heterozygotes will develop AD and will do so at a mean age of 73, while 9% of noncarriers will develop the disease and are usually diagnosed around age 80.

As in the larger cohort, subjects in the substudy took the battery of neuropsychological tests every 2 years. The battery consists of four tests in each of five domains: executive, memory, language, spatial, and behavioral. Significant decline was defined as a drop of two standard deviations beyond that of the entire cohort in one or more domain test scores. Subjects were judged to have cognitive domain decline if their scores were lower on at least two tests in any single domain.

“We found that there was really no difference between the genetic subgroups in patterns of decline in the younger group of patients, aged 50–59 years,” Dr. Caselli said. “Some had no decline, some showed improvement, and some had domain decline, but there were no statistically significant differences.”

Significant differences did emerge in the group of 60- to 69-year-olds, however. Homozygotes had the highest proportion of cognitive decline, with 40% showing domain decline, compared with 8% of heterozygotes and noncarriers. None of the older noncarriers or heterozygotes experienced a decline in two or more domains, while this occurred in 20% of the homozygous subjects.

Dr. Caselli has additional data on 97 subjects who have been tested again in the years following their initial decline. “We saw that it was those who initially declined on memory who tended to continue to decline significantly in other areas, and if that subsequent decline was in the memory domain, it was even more pronounced.”

Seven subjects developed MCI or AD during the study; five of these were apo ϵ4 homozygotes, one was an apo ϵ4 heterozygote, and one was a noncarrier. “Typically, it took about 2 years following the epoch of domain decline for the diagnosis to occur.”

By looking at the larger cohort and including subjects aged 30–90 years, Dr. Caselli also found a striking age-related separation of memory domain performance between apo ϵ4 carriers and noncarriers. From age 30 to 60, memory performance on the Auditory Verbal Learning Test, which requires subjects to recall 15 words from a list, declined slowly and consistently, from a mean of 11 words at age 30 to about 9.5 words at age 60. Immediately thereafter, however, the groups separated. Noncarriers continued a slow, almost linear decline, and by age 90, their predicted mean word recall was about 8.5 of 15. But carriers entered a phase of sharper decline; by the time they reached 90 years, their predicted mean word recall was about 5.5.

The slope of decline exhibited by the noncarriers represents normal, age-related memory loss, Dr. Caselli said. The sharper post-60-year decline in apo ϵ4 carriers probably reflects a direct or indirect effect of the ϵ4 gene.

But although these cognitive changes appeared mainly after age 60, imaging studies on some of the younger carriers suggest that some brain areas may be vulnerable much earlier in life. PET scanning of presymptomatic 50- to 59-year-old homozygotes showed areas of decreased glucose metabolism in brain regions associated with Alzheimer's disease pathology: the posterior cingulate gyrus, parietal and temporal lobes, and prefrontal cortex. PET scans of 20- to 39-year-olds with one copy of the allele showed similar, although smaller, areas of decreased metabolism.

“So what does that mean?” asked Dr. Caselli. “If you look at all the work out there–the Nun Study, brain imaging, and pathology studies of apo ϵ4 carriers–you get the idea that little pieces of AD pathology happen throughout young adult life, but we don't see young people developing progressive dementia unless there's an autosomal dominant mutation. The fact is, we don't know whether these early changes reflect a sort of nonprogressive pathology or some basic biologic vulnerability that marks the territory of later decline.”

There are plenty of theories about the possible connection between apo ϵ4 status and Alzheimer's pathology, Dr. Caselli said. Most focus on the pathologic function of the apo ϵ4 isoform. Research from the 1990s suggests that it enhances amyloid deposition, reduces neurite outgrowth and protection against oxidative stress, and cuts the efficiency of neuronal and synaptic repair. Most recently, researchers at the University of California, San Francisco, have suggested that the apo ϵ4 isoform can generate a cytotoxic carboxyl fragment. This truncated form of the protein is thought to induce neuronal inclusions that are similar to neurofibrillatory tangles, containing phosphorylated tau and high-molecular-weight neurofilaments (Proc. Natl. Acad. Sci. USA 2001;98:8838–43).

“The science on this is pretty well established, but whether it's key to AD pathogenesis is still undergoing further study,” Dr. Caselli said. Another recently proposed connection is the relationship between apo ϵ4 status and the demyelination in the frontal lobe and corpus callosum, Dr. Caselli noted. These brain regions, which continue to lay down myelin until middle age, also appear most susceptible to myelin breakdown, wrote Dr. George Bartzokis, director of the UCLA Memory Disorders and Alzheimer's Disease Clinic in Los Angeles.

Dr. Bartzokis's study of 104 healthy subjects aged 75 years and younger found that those with the apo ϵ4 genotype had the highest level of demyelination in frontal lobe white matter and the genu of the corpus callosum. The apo ϵ2 genotype appeared protective of demyelination, while those who were apo ϵ3 positive had an intermediate level of demyelination.

The connection may be the dearth of apo ϵ molecules in the ϵ4 genotypes. Apo ϵ helps maintain neuronal health by degrading damaged myelin and recycling the lipids for rapid repair. Those who are apo ϵ2 positive have the highest number of apo ϵ molecules available for this constant repair process; those who are apo ϵ4 positive have the lowest number, while apo ϵ3-positive subjects have an intermediate number (Arch. Gen. Psychiatry 2006;63:63–72; Proc. Natl. Acad. Sci. USA 2006;103:5641–3).

Age 60 seems to be the defining year for many homozygous carriers of the apolipoprotein ϵ4 gene –the time when age-related changes in cognition focus more on memory and begin a steeper decline into mild cognitive impairment and, eventually Alzheimer's disease, according to new unpublished observations from a longitudinal study of apo ϵ4 carriers and normal controls.

“We saw normal age-related patterns of memory loss that occurred before age 60, but during this period, we didn't see any significant cognitive differences between the apo ϵ4 homozygotes, heterozygotes, and noncarriers,” said Dr. Richard Caselli, a lead investigator for the Arizona Apoϵ4 Cohort Longitudinal Study of Cognitively Normal Individuals. “But our latest information shows that at around age 60, a separation begins and continues for as long as we have been able to follow our subjects. There is a particular pattern of decline in apo ϵ4 homozygotes that tends to precede any diagnosis of mild cognitive impairment [MCI] or anything that can be seen with routine clinical brain imaging,” he said in an interview.

This pattern suggests that pathologic changes could be occurring earlier in homozygous ϵ4 carriers, although the exact nature of these changes, and their triggers, remain speculative, said Dr. Caselli, chairman of the department of neurology at the Mayo Clinic, Scottsdale, Ariz., and a member of the Arizona Alzheimer's Disease Consortium.

The Arizona cohort was initiated in 1994, and now includes more than 600 people, enrolled at ages 20–90 years, who have at least one first-degree relative with Alzheimer's disease. The subjects are all genotyped for the apo ϵ4 allele, and undergo extensive neuropsychological testing every 2 years.

Dr. Caselli's recent substudy focuses on 214 of these subjects aged 50–69 years. Almost half are apo ϵ4 carriers–43 homozygous and 59 heterozygous.

The study set out to characterize the effect of apo ϵ4 status on the development of presymptomatic cognitive changes. It's well known that the gene has a dose-response effect on the age at AD diagnosis: 80%–90% of homozygotes will develop the disease, at a mean age of 68 years. About 30% of heterozygotes will develop AD and will do so at a mean age of 73, while 9% of noncarriers will develop the disease and are usually diagnosed around age 80.

As in the larger cohort, subjects in the substudy took the battery of neuropsychological tests every 2 years. The battery consists of four tests in each of five domains: executive, memory, language, spatial, and behavioral. Significant decline was defined as a drop of two standard deviations beyond that of the entire cohort in one or more domain test scores. Subjects were judged to have cognitive domain decline if their scores were lower on at least two tests in any single domain.

“We found that there was really no difference between the genetic subgroups in patterns of decline in the younger group of patients, aged 50–59 years,” Dr. Caselli said. “Some had no decline, some showed improvement, and some had domain decline, but there were no statistically significant differences.”

Significant differences did emerge in the group of 60- to 69-year-olds, however. Homozygotes had the highest proportion of cognitive decline, with 40% showing domain decline, compared with 8% of heterozygotes and noncarriers. None of the older noncarriers or heterozygotes experienced a decline in two or more domains, while this occurred in 20% of the homozygous subjects.

Dr. Caselli has additional data on 97 subjects who have been tested again in the years following their initial decline. “We saw that it was those who initially declined on memory who tended to continue to decline significantly in other areas, and if that subsequent decline was in the memory domain, it was even more pronounced.”

Seven subjects developed MCI or AD during the study; five of these were apo ϵ4 homozygotes, one was an apo ϵ4 heterozygote, and one was a noncarrier. “Typically, it took about 2 years following the epoch of domain decline for the diagnosis to occur.”

By looking at the larger cohort and including subjects aged 30–90 years, Dr. Caselli also found a striking age-related separation of memory domain performance between apo ϵ4 carriers and noncarriers. From age 30 to 60, memory performance on the Auditory Verbal Learning Test, which requires subjects to recall 15 words from a list, declined slowly and consistently, from a mean of 11 words at age 30 to about 9.5 words at age 60. Immediately thereafter, however, the groups separated. Noncarriers continued a slow, almost linear decline, and by age 90, their predicted mean word recall was about 8.5 of 15. But carriers entered a phase of sharper decline; by the time they reached 90 years, their predicted mean word recall was about 5.5.

The slope of decline exhibited by the noncarriers represents normal, age-related memory loss, Dr. Caselli said. The sharper post-60-year decline in apo ϵ4 carriers probably reflects a direct or indirect effect of the ϵ4 gene.

But although these cognitive changes appeared mainly after age 60, imaging studies on some of the younger carriers suggest that some brain areas may be vulnerable much earlier in life. PET scanning of presymptomatic 50- to 59-year-old homozygotes showed areas of decreased glucose metabolism in brain regions associated with Alzheimer's disease pathology: the posterior cingulate gyrus, parietal and temporal lobes, and prefrontal cortex. PET scans of 20- to 39-year-olds with one copy of the allele showed similar, although smaller, areas of decreased metabolism.

“So what does that mean?” asked Dr. Caselli. “If you look at all the work out there–the Nun Study, brain imaging, and pathology studies of apo ϵ4 carriers–you get the idea that little pieces of AD pathology happen throughout young adult life, but we don't see young people developing progressive dementia unless there's an autosomal dominant mutation. The fact is, we don't know whether these early changes reflect a sort of nonprogressive pathology or some basic biologic vulnerability that marks the territory of later decline.”

There are plenty of theories about the possible connection between apo ϵ4 status and Alzheimer's pathology, Dr. Caselli said. Most focus on the pathologic function of the apo ϵ4 isoform. Research from the 1990s suggests that it enhances amyloid deposition, reduces neurite outgrowth and protection against oxidative stress, and cuts the efficiency of neuronal and synaptic repair. Most recently, researchers at the University of California, San Francisco, have suggested that the apo ϵ4 isoform can generate a cytotoxic carboxyl fragment. This truncated form of the protein is thought to induce neuronal inclusions that are similar to neurofibrillatory tangles, containing phosphorylated tau and high-molecular-weight neurofilaments (Proc. Natl. Acad. Sci. USA 2001;98:8838–43).

“The science on this is pretty well established, but whether it's key to AD pathogenesis is still undergoing further study,” Dr. Caselli said. Another recently proposed connection is the relationship between apo ϵ4 status and the demyelination in the frontal lobe and corpus callosum, Dr. Caselli noted. These brain regions, which continue to lay down myelin until middle age, also appear most susceptible to myelin breakdown, wrote Dr. George Bartzokis, director of the UCLA Memory Disorders and Alzheimer's Disease Clinic in Los Angeles.

Dr. Bartzokis's study of 104 healthy subjects aged 75 years and younger found that those with the apo ϵ4 genotype had the highest level of demyelination in frontal lobe white matter and the genu of the corpus callosum. The apo ϵ2 genotype appeared protective of demyelination, while those who were apo ϵ3 positive had an intermediate level of demyelination.

The connection may be the dearth of apo ϵ molecules in the ϵ4 genotypes. Apo ϵ helps maintain neuronal health by degrading damaged myelin and recycling the lipids for rapid repair. Those who are apo ϵ2 positive have the highest number of apo ϵ molecules available for this constant repair process; those who are apo ϵ4 positive have the lowest number, while apo ϵ3-positive subjects have an intermediate number (Arch. Gen. Psychiatry 2006;63:63–72; Proc. Natl. Acad. Sci. USA 2006;103:5641–3).

Age 60 seems to be the defining year for many homozygous carriers of the apolipoprotein ϵ4 gene –the time when age-related changes in cognition focus more on memory and begin a steeper decline into mild cognitive impairment and, eventually Alzheimer's disease, according to new unpublished observations from a longitudinal study of apo ϵ4 carriers and normal controls.

“We saw normal age-related patterns of memory loss that occurred before age 60, but during this period, we didn't see any significant cognitive differences between the apo ϵ4 homozygotes, heterozygotes, and noncarriers,” said Dr. Richard Caselli, a lead investigator for the Arizona Apoϵ4 Cohort Longitudinal Study of Cognitively Normal Individuals. “But our latest information shows that at around age 60, a separation begins and continues for as long as we have been able to follow our subjects. There is a particular pattern of decline in apo ϵ4 homozygotes that tends to precede any diagnosis of mild cognitive impairment [MCI] or anything that can be seen with routine clinical brain imaging,” he said in an interview.

This pattern suggests that pathologic changes could be occurring earlier in homozygous ϵ4 carriers, although the exact nature of these changes, and their triggers, remain speculative, said Dr. Caselli, chairman of the department of neurology at the Mayo Clinic, Scottsdale, Ariz., and a member of the Arizona Alzheimer's Disease Consortium.

The Arizona cohort was initiated in 1994, and now includes more than 600 people, enrolled at ages 20–90 years, who have at least one first-degree relative with Alzheimer's disease. The subjects are all genotyped for the apo ϵ4 allele, and undergo extensive neuropsychological testing every 2 years.

Dr. Caselli's recent substudy focuses on 214 of these subjects aged 50–69 years. Almost half are apo ϵ4 carriers–43 homozygous and 59 heterozygous.

The study set out to characterize the effect of apo ϵ4 status on the development of presymptomatic cognitive changes. It's well known that the gene has a dose-response effect on the age at AD diagnosis: 80%–90% of homozygotes will develop the disease, at a mean age of 68 years. About 30% of heterozygotes will develop AD and will do so at a mean age of 73, while 9% of noncarriers will develop the disease and are usually diagnosed around age 80.

As in the larger cohort, subjects in the substudy took the battery of neuropsychological tests every 2 years. The battery consists of four tests in each of five domains: executive, memory, language, spatial, and behavioral. Significant decline was defined as a drop of two standard deviations beyond that of the entire cohort in one or more domain test scores. Subjects were judged to have cognitive domain decline if their scores were lower on at least two tests in any single domain.

“We found that there was really no difference between the genetic subgroups in patterns of decline in the younger group of patients, aged 50–59 years,” Dr. Caselli said. “Some had no decline, some showed improvement, and some had domain decline, but there were no statistically significant differences.”

Significant differences did emerge in the group of 60- to 69-year-olds, however. Homozygotes had the highest proportion of cognitive decline, with 40% showing domain decline, compared with 8% of heterozygotes and noncarriers. None of the older noncarriers or heterozygotes experienced a decline in two or more domains, while this occurred in 20% of the homozygous subjects.

Dr. Caselli has additional data on 97 subjects who have been tested again in the years following their initial decline. “We saw that it was those who initially declined on memory who tended to continue to decline significantly in other areas, and if that subsequent decline was in the memory domain, it was even more pronounced.”

Seven subjects developed MCI or AD during the study; five of these were apo ϵ4 homozygotes, one was an apo ϵ4 heterozygote, and one was a noncarrier. “Typically, it took about 2 years following the epoch of domain decline for the diagnosis to occur.”

By looking at the larger cohort and including subjects aged 30–90 years, Dr. Caselli also found a striking age-related separation of memory domain performance between apo ϵ4 carriers and noncarriers. From age 30 to 60, memory performance on the Auditory Verbal Learning Test, which requires subjects to recall 15 words from a list, declined slowly and consistently, from a mean of 11 words at age 30 to about 9.5 words at age 60. Immediately thereafter, however, the groups separated. Noncarriers continued a slow, almost linear decline, and by age 90, their predicted mean word recall was about 8.5 of 15. But carriers entered a phase of sharper decline; by the time they reached 90 years, their predicted mean word recall was about 5.5.

The slope of decline exhibited by the noncarriers represents normal, age-related memory loss, Dr. Caselli said. The sharper post-60-year decline in apo ϵ4 carriers probably reflects a direct or indirect effect of the ϵ4 gene.

But although these cognitive changes appeared mainly after age 60, imaging studies on some of the younger carriers suggest that some brain areas may be vulnerable much earlier in life. PET scanning of presymptomatic 50- to 59-year-old homozygotes showed areas of decreased glucose metabolism in brain regions associated with Alzheimer's disease pathology: the posterior cingulate gyrus, parietal and temporal lobes, and prefrontal cortex. PET scans of 20- to 39-year-olds with one copy of the allele showed similar, although smaller, areas of decreased metabolism.

“So what does that mean?” asked Dr. Caselli. “If you look at all the work out there–the Nun Study, brain imaging, and pathology studies of apo ϵ4 carriers–you get the idea that little pieces of AD pathology happen throughout young adult life, but we don't see young people developing progressive dementia unless there's an autosomal dominant mutation. The fact is, we don't know whether these early changes reflect a sort of nonprogressive pathology or some basic biologic vulnerability that marks the territory of later decline.”

There are plenty of theories about the possible connection between apo ϵ4 status and Alzheimer's pathology, Dr. Caselli said. Most focus on the pathologic function of the apo ϵ4 isoform. Research from the 1990s suggests that it enhances amyloid deposition, reduces neurite outgrowth and protection against oxidative stress, and cuts the efficiency of neuronal and synaptic repair. Most recently, researchers at the University of California, San Francisco, have suggested that the apo ϵ4 isoform can generate a cytotoxic carboxyl fragment. This truncated form of the protein is thought to induce neuronal inclusions that are similar to neurofibrillatory tangles, containing phosphorylated tau and high-molecular-weight neurofilaments (Proc. Natl. Acad. Sci. USA 2001;98:8838–43).

“The science on this is pretty well established, but whether it's key to AD pathogenesis is still undergoing further study,” Dr. Caselli said. Another recently proposed connection is the relationship between apo ϵ4 status and the demyelination in the frontal lobe and corpus callosum, Dr. Caselli noted. These brain regions, which continue to lay down myelin until middle age, also appear most susceptible to myelin breakdown, wrote Dr. George Bartzokis, director of the UCLA Memory Disorders and Alzheimer's Disease Clinic in Los Angeles.

Dr. Bartzokis's study of 104 healthy subjects aged 75 years and younger found that those with the apo ϵ4 genotype had the highest level of demyelination in frontal lobe white matter and the genu of the corpus callosum. The apo ϵ2 genotype appeared protective of demyelination, while those who were apo ϵ3 positive had an intermediate level of demyelination.

The connection may be the dearth of apo ϵ molecules in the ϵ4 genotypes. Apo ϵ helps maintain neuronal health by degrading damaged myelin and recycling the lipids for rapid repair. Those who are apo ϵ2 positive have the highest number of apo ϵ molecules available for this constant repair process; those who are apo ϵ4 positive have the lowest number, while apo ϵ3-positive subjects have an intermediate number (Arch. Gen. Psychiatry 2006;63:63–72; Proc. Natl. Acad. Sci. USA 2006;103:5641–3).

A $60 million, 6-year study is being launched to find and validate biologic and imaging markers that could be used as objective measures of therapeutic response in Alzheimer's disease.

The results of the Alzheimer's Disease Neuroimaging Initiative (ADNI) could dramatically shorten clinical trials of potential therapies by sidestepping the years of cognitive testing now necessary to determine a drug effects, according to experts interviewed for this article.

The time is ripe for an objective biologic marker of disease progression. Several potentially disease-modifying drugs are in phase II trials, with cognitive measures the only validated treatment outcomes. Relying solely on cognition to determine treatment effect is problematic in many ways, said Dr. Michael W. Weiner, ADNI's principal investigator and the director of the Veterans Administration's Center for Imaging of Neurodegenerative Disease, San Francisco.

Everyone with AD declines, but they don't do so in a linear fashion, Dr. Weiner said “There is a lot of variability in these kinds of measures. One day, a patient might do well, and the next, do poorly depending on his general health, emotional status, or a number of other factors. This variability really affects the ability to determine a true treatment effect.”

Nor, using cognitive outcomes alone, is it possible to distinguish between a drug's effects on disease progression and any memory-enhancing effects it also might exhibit, he said.

A validated biomarker, on the other hand, could show a drug's true effect, with profound influence on drug development, said Dr. John Q. Trojanowski, director of the Alzheimer's Disease Center at the University of Pennsylvania, Philadelphia, and leader of ADNI's biomarker core. “The pace of drug discovery would quicken incredibly, and the costs come down incredibly, if we had a chemical or imaging marker that reflected reversing or blocking disease progression.”

Funded by a mix of federal and private sources, ADNI will search for such markers in 800 patients aged 55–90 years: 200 with AD, 400 with mild cognitive impairment (MCI), and 200 healthy controls. Examining three imaging techniques and four biomarkers, the study aims to find predictors of progression in AD patients, and predictors of transformation from normal to MCI and from MCI to AD.

Since the study's main goal is to improve the climate for clinical trials, pharmaceutical companies have an enormous stake in its outcome. Faced with a cost of up to $1 billion to bring just one drug to market, it's not surprising that 13 of the world's largest drug makers have agreed to fund about a third of ADNI's cost. The National Institutes of Health is footing the rest of the bill.

“For an investment of a few million dollars, [pharmaceutical companies are] hoping for biomarkers that would enable them to bring the cost of their clinical trials down by millions of dollars,” said Dr. Trojanowski. “Chances are that this is a reasonable expectation.”

Patient enrollment wrapped up last year, and now the work is beginning at 57 centers across the United States and Canada. In addition to the baseline visit, AD patients will have three follow-ups (6, 12, and 24 months). Normal controls will have four follow-ups, including an additional visit at 36 months. MCI patients will be seen a total of six times: at baseline and at 6, 12, 18, 24, and 36 months.

Each cohort will have apolipoprotein E genotyping at baseline and undergo standard magnetic resonance imaging at all time points. Half also will receive fluorodeoxyglucose PET scanning at each time point, and another 25% will undergo the more sensitive 3-Tesla MRI, which allows metabolic and physiologic imaging.

A group of 120 patients also will be enrolled in a substudy of Pittsburgh compound B PET scanning. Because the compound binds to amyloid plaques in the brain, it offers a reliable way by which to trace disease progression and may be able to detect early pathologic changes before cognitive changes develop.

At all follow-up visits, patients will donate blood and urine for evaluation for these potential markers; 55% will undergo at least two lumbar punctures for the collection of cerebrospinal fluid. ADNI will focus on four of the most promising biomarkers:

▸ Homocysteine, while not diagnostically significant, may reflect disease progression.

▸ Isoprostanes, indicators of oxidative damage, have been shown to be increased in the hippocampi and CSF of AD patients.

▸ Phosphorylated tau and amyloid-β are the neuropathologic hallmarks of AD tangles and plaques, respectively.

Most research indicates that patients with low levels of amyloid-β and high levels of tau in CSF are more likely to have AD, although some recent studies have challenged this idea.

In fact, none of the biomarkers or imaging modalities included in the ADNI study has been validated in large numbers of patients. Without validation, none can be used as a primary outcome in drug trials, leaving researchers to fall back on imprecise cognitive measures for their main assessment of efficacy, Dr. Trojanowski said.

“If, instead, we could show that biomarkers were changing during a drug trial, in the same way that we can show changes in cholesterol or blood sugar in response to drugs, then we could have a very, very powerful new tool,” Dr. Trojanowski said.

ADNI is an important study in a worldwide effort to discover clinical markers of AD, the investigators agreed. Australia, Japan, and some European countries are undertaking similar projects. Termed “World Wide ADNI,” this informal network will facilitate the performance of international treatment trials and, ultimately, the approval of disease-modifying treatments worldwide. Researchers from these studies are working to harmonize the methodologies, increasing the statistical power of their combined results, Dr. Trojanowski said.

Dr. Weiner noted before ADNI has evaluated even a single patient, it has already changed the way scientists are attacking AD research. Most AD biomarker and imaging research now consists of small studies, each with a unique methodology. Researchers hold their information close to the vest until publication, with months and sometimes years elapsing between the study's conclusion and the sharing of its results.

Not only will all 57 ADNI study centers follow consistent protocols for all the imaging studies and biomarker collection, but also those protocols will be available to any researcher pursuing an independent study. The sharing of methodology and, eventually, of results, is another of ADNI's unique characteristics, Dr. Weiner said. All data will be sent to centralized storage hubs, which will be freely available to anyone–researcher, physician, patient, or family member–who applies for access. The information won't be subject to embargo; scans and biomarker measurements will be available online as soon as they're processed.

“We are going to share everything we get in an unprecedented way. Allowing other researchers to have immediate access to the data is going to maximize this study's effect.”

For more information, including filing a request for access to eventual data, physicians should go to www.adni-info.org

Availability of Markers Bound to Raise Thorny Questions

The first benefits of biologic or imaging markers would be felt in research, where they could hasten the development of disease-modifying drugs. But once those drugs are available, such markers also will be used to identify people most likely to benefit from them and could become part of screening to identify those at risk of developing Alzheimer's disease years before symptoms emerge.

“If and when these therapies become available, who will get them?” asked Dr. Ronald Petersen, director of the Mayo Clinical Alzheimer's Disease Research Center, Rochester, Minn. “Are we going to give them to people with memory impairment first? Or ultimately to those who are asymptomatic but who are at risk? And how will we know who these people are?”

Dr. Petersen hopes eventual findings from his own study help answer some of these questions. He is the lead investigator for the Mayo Clinic Study of Aging, a longitudinal study of 2,000 people aged 70–89 years. The National Institute on Aging provides the study's $1.5 million annual funding package.

In addition to characterizing what brain aging looks like, the study will examine the influence of genetics, family history, and medical comorbidities on the risk of developing Alzheimer's. It will combine this information with cognitive testing, biomarkers, and different imaging modalities in an effort to construct a multivariate model that could predict which apparently healthy people will develop mild cognitive impairment and who might then progress to Alzheimer's.

The Clinical Study of Aging is not related to Alzheimer's Disease Neuroimaging Initiative (ADNI), but both aim to identify the most informative biomarkers and imaging techniques for AD. Eventually, Dr. Petersen predicted, data from both studies might be used to construct a layered screening technique that could identify those who would benefit from early disease prevention or disease-modifying therapy.

Such a model probably would progress from least- to most-invasive testing depending on individual risk, he said. Patients flagged as moderate or high risk might then receive a functional brain scan and a lumbar puncture for cerebrospinal fluid marker sampling. Amyloid imaging with Pittsburgh compound B might also be part of the work-up.

In stratifying patients for treatment, “Earlier may be better than later,” said Dr. John Morris, director of the Alzheimer's Disease Research Center at Washington University, St. Louis. “We know that by the time people start experiencing memory problems, the neuropathological lesions of AD [plaques and tangles] already are present in the brain and, in brain regions vulnerable to AD, nerve cells and their synapses have been lost. Even if truly effective drugs are developed, giving them to people with symptomatic AD–even to those with mild cognitive impairment–may be too late because, by the time symptoms appear, the brain already has been damaged.”

ADNI does not include what could be the critical population in this theory–younger normal controls. “While the study does include nondemented patients (200 normal individuals and 400 with MCI), these people are all older than 55. It is still speculative to consider when the AD process begins in the brain, but it may be in midlife or even earlier.”

To be truly effective in combating the disease, he suggested, it will be important to study younger groups to detect the beginnings of AD with imaging or biological markers before symptoms appear so that eventually therapies hopefully can be given to prevent the occurrence of dementia.

“It's similar to treating atherosclerosis,” he said. “It's a lot easier to treat cholesterol levels to prevent vascular damage than to wait until a heart attack occurs and try to fix things from that point.”

A $60 million, 6-year study is being launched to find and validate biologic and imaging markers that could be used as objective measures of therapeutic response in Alzheimer's disease.

The results of the Alzheimer's Disease Neuroimaging Initiative (ADNI) could dramatically shorten clinical trials of potential therapies by sidestepping the years of cognitive testing now necessary to determine a drug effects, according to experts interviewed for this article.

The time is ripe for an objective biologic marker of disease progression. Several potentially disease-modifying drugs are in phase II trials, with cognitive measures the only validated treatment outcomes. Relying solely on cognition to determine treatment effect is problematic in many ways, said Dr. Michael W. Weiner, ADNI's principal investigator and the director of the Veterans Administration's Center for Imaging of Neurodegenerative Disease, San Francisco.

Everyone with AD declines, but they don't do so in a linear fashion, Dr. Weiner said “There is a lot of variability in these kinds of measures. One day, a patient might do well, and the next, do poorly depending on his general health, emotional status, or a number of other factors. This variability really affects the ability to determine a true treatment effect.”

Nor, using cognitive outcomes alone, is it possible to distinguish between a drug's effects on disease progression and any memory-enhancing effects it also might exhibit, he said.

A validated biomarker, on the other hand, could show a drug's true effect, with profound influence on drug development, said Dr. John Q. Trojanowski, director of the Alzheimer's Disease Center at the University of Pennsylvania, Philadelphia, and leader of ADNI's biomarker core. “The pace of drug discovery would quicken incredibly, and the costs come down incredibly, if we had a chemical or imaging marker that reflected reversing or blocking disease progression.”

Funded by a mix of federal and private sources, ADNI will search for such markers in 800 patients aged 55–90 years: 200 with AD, 400 with mild cognitive impairment (MCI), and 200 healthy controls. Examining three imaging techniques and four biomarkers, the study aims to find predictors of progression in AD patients, and predictors of transformation from normal to MCI and from MCI to AD.

Since the study's main goal is to improve the climate for clinical trials, pharmaceutical companies have an enormous stake in its outcome. Faced with a cost of up to $1 billion to bring just one drug to market, it's not surprising that 13 of the world's largest drug makers have agreed to fund about a third of ADNI's cost. The National Institutes of Health is footing the rest of the bill.

“For an investment of a few million dollars, [pharmaceutical companies are] hoping for biomarkers that would enable them to bring the cost of their clinical trials down by millions of dollars,” said Dr. Trojanowski. “Chances are that this is a reasonable expectation.”

Patient enrollment wrapped up last year, and now the work is beginning at 57 centers across the United States and Canada. In addition to the baseline visit, AD patients will have three follow-ups (6, 12, and 24 months). Normal controls will have four follow-ups, including an additional visit at 36 months. MCI patients will be seen a total of six times: at baseline and at 6, 12, 18, 24, and 36 months.

Each cohort will have apolipoprotein E genotyping at baseline and undergo standard magnetic resonance imaging at all time points. Half also will receive fluorodeoxyglucose PET scanning at each time point, and another 25% will undergo the more sensitive 3-Tesla MRI, which allows metabolic and physiologic imaging.

A group of 120 patients also will be enrolled in a substudy of Pittsburgh compound B PET scanning. Because the compound binds to amyloid plaques in the brain, it offers a reliable way by which to trace disease progression and may be able to detect early pathologic changes before cognitive changes develop.

At all follow-up visits, patients will donate blood and urine for evaluation for these potential markers; 55% will undergo at least two lumbar punctures for the collection of cerebrospinal fluid. ADNI will focus on four of the most promising biomarkers:

▸ Homocysteine, while not diagnostically significant, may reflect disease progression.

▸ Isoprostanes, indicators of oxidative damage, have been shown to be increased in the hippocampi and CSF of AD patients.

▸ Phosphorylated tau and amyloid-β are the neuropathologic hallmarks of AD tangles and plaques, respectively.

Most research indicates that patients with low levels of amyloid-β and high levels of tau in CSF are more likely to have AD, although some recent studies have challenged this idea.

In fact, none of the biomarkers or imaging modalities included in the ADNI study has been validated in large numbers of patients. Without validation, none can be used as a primary outcome in drug trials, leaving researchers to fall back on imprecise cognitive measures for their main assessment of efficacy, Dr. Trojanowski said.

“If, instead, we could show that biomarkers were changing during a drug trial, in the same way that we can show changes in cholesterol or blood sugar in response to drugs, then we could have a very, very powerful new tool,” Dr. Trojanowski said.

ADNI is an important study in a worldwide effort to discover clinical markers of AD, the investigators agreed. Australia, Japan, and some European countries are undertaking similar projects. Termed “World Wide ADNI,” this informal network will facilitate the performance of international treatment trials and, ultimately, the approval of disease-modifying treatments worldwide. Researchers from these studies are working to harmonize the methodologies, increasing the statistical power of their combined results, Dr. Trojanowski said.

Dr. Weiner noted before ADNI has evaluated even a single patient, it has already changed the way scientists are attacking AD research. Most AD biomarker and imaging research now consists of small studies, each with a unique methodology. Researchers hold their information close to the vest until publication, with months and sometimes years elapsing between the study's conclusion and the sharing of its results.

Not only will all 57 ADNI study centers follow consistent protocols for all the imaging studies and biomarker collection, but also those protocols will be available to any researcher pursuing an independent study. The sharing of methodology and, eventually, of results, is another of ADNI's unique characteristics, Dr. Weiner said. All data will be sent to centralized storage hubs, which will be freely available to anyone–researcher, physician, patient, or family member–who applies for access. The information won't be subject to embargo; scans and biomarker measurements will be available online as soon as they're processed.

“We are going to share everything we get in an unprecedented way. Allowing other researchers to have immediate access to the data is going to maximize this study's effect.”

For more information, including filing a request for access to eventual data, physicians should go to www.adni-info.org

Availability of Markers Bound to Raise Thorny Questions

The first benefits of biologic or imaging markers would be felt in research, where they could hasten the development of disease-modifying drugs. But once those drugs are available, such markers also will be used to identify people most likely to benefit from them and could become part of screening to identify those at risk of developing Alzheimer's disease years before symptoms emerge.

“If and when these therapies become available, who will get them?” asked Dr. Ronald Petersen, director of the Mayo Clinical Alzheimer's Disease Research Center, Rochester, Minn. “Are we going to give them to people with memory impairment first? Or ultimately to those who are asymptomatic but who are at risk? And how will we know who these people are?”

Dr. Petersen hopes eventual findings from his own study help answer some of these questions. He is the lead investigator for the Mayo Clinic Study of Aging, a longitudinal study of 2,000 people aged 70–89 years. The National Institute on Aging provides the study's $1.5 million annual funding package.

In addition to characterizing what brain aging looks like, the study will examine the influence of genetics, family history, and medical comorbidities on the risk of developing Alzheimer's. It will combine this information with cognitive testing, biomarkers, and different imaging modalities in an effort to construct a multivariate model that could predict which apparently healthy people will develop mild cognitive impairment and who might then progress to Alzheimer's.

The Clinical Study of Aging is not related to Alzheimer's Disease Neuroimaging Initiative (ADNI), but both aim to identify the most informative biomarkers and imaging techniques for AD. Eventually, Dr. Petersen predicted, data from both studies might be used to construct a layered screening technique that could identify those who would benefit from early disease prevention or disease-modifying therapy.

Such a model probably would progress from least- to most-invasive testing depending on individual risk, he said. Patients flagged as moderate or high risk might then receive a functional brain scan and a lumbar puncture for cerebrospinal fluid marker sampling. Amyloid imaging with Pittsburgh compound B might also be part of the work-up.

In stratifying patients for treatment, “Earlier may be better than later,” said Dr. John Morris, director of the Alzheimer's Disease Research Center at Washington University, St. Louis. “We know that by the time people start experiencing memory problems, the neuropathological lesions of AD [plaques and tangles] already are present in the brain and, in brain regions vulnerable to AD, nerve cells and their synapses have been lost. Even if truly effective drugs are developed, giving them to people with symptomatic AD–even to those with mild cognitive impairment–may be too late because, by the time symptoms appear, the brain already has been damaged.”

ADNI does not include what could be the critical population in this theory–younger normal controls. “While the study does include nondemented patients (200 normal individuals and 400 with MCI), these people are all older than 55. It is still speculative to consider when the AD process begins in the brain, but it may be in midlife or even earlier.”

To be truly effective in combating the disease, he suggested, it will be important to study younger groups to detect the beginnings of AD with imaging or biological markers before symptoms appear so that eventually therapies hopefully can be given to prevent the occurrence of dementia.

“It's similar to treating atherosclerosis,” he said. “It's a lot easier to treat cholesterol levels to prevent vascular damage than to wait until a heart attack occurs and try to fix things from that point.”

A $60 million, 6-year study is being launched to find and validate biologic and imaging markers that could be used as objective measures of therapeutic response in Alzheimer's disease.

The results of the Alzheimer's Disease Neuroimaging Initiative (ADNI) could dramatically shorten clinical trials of potential therapies by sidestepping the years of cognitive testing now necessary to determine a drug effects, according to experts interviewed for this article.

The time is ripe for an objective biologic marker of disease progression. Several potentially disease-modifying drugs are in phase II trials, with cognitive measures the only validated treatment outcomes. Relying solely on cognition to determine treatment effect is problematic in many ways, said Dr. Michael W. Weiner, ADNI's principal investigator and the director of the Veterans Administration's Center for Imaging of Neurodegenerative Disease, San Francisco.

Everyone with AD declines, but they don't do so in a linear fashion, Dr. Weiner said “There is a lot of variability in these kinds of measures. One day, a patient might do well, and the next, do poorly depending on his general health, emotional status, or a number of other factors. This variability really affects the ability to determine a true treatment effect.”

Nor, using cognitive outcomes alone, is it possible to distinguish between a drug's effects on disease progression and any memory-enhancing effects it also might exhibit, he said.

A validated biomarker, on the other hand, could show a drug's true effect, with profound influence on drug development, said Dr. John Q. Trojanowski, director of the Alzheimer's Disease Center at the University of Pennsylvania, Philadelphia, and leader of ADNI's biomarker core. “The pace of drug discovery would quicken incredibly, and the costs come down incredibly, if we had a chemical or imaging marker that reflected reversing or blocking disease progression.”

Funded by a mix of federal and private sources, ADNI will search for such markers in 800 patients aged 55–90 years: 200 with AD, 400 with mild cognitive impairment (MCI), and 200 healthy controls. Examining three imaging techniques and four biomarkers, the study aims to find predictors of progression in AD patients, and predictors of transformation from normal to MCI and from MCI to AD.

Since the study's main goal is to improve the climate for clinical trials, pharmaceutical companies have an enormous stake in its outcome. Faced with a cost of up to $1 billion to bring just one drug to market, it's not surprising that 13 of the world's largest drug makers have agreed to fund about a third of ADNI's cost. The National Institutes of Health is footing the rest of the bill.

“For an investment of a few million dollars, [pharmaceutical companies are] hoping for biomarkers that would enable them to bring the cost of their clinical trials down by millions of dollars,” said Dr. Trojanowski. “Chances are that this is a reasonable expectation.”

Patient enrollment wrapped up last year, and now the work is beginning at 57 centers across the United States and Canada. In addition to the baseline visit, AD patients will have three follow-ups (6, 12, and 24 months). Normal controls will have four follow-ups, including an additional visit at 36 months. MCI patients will be seen a total of six times: at baseline and at 6, 12, 18, 24, and 36 months.

Each cohort will have apolipoprotein E genotyping at baseline and undergo standard magnetic resonance imaging at all time points. Half also will receive fluorodeoxyglucose PET scanning at each time point, and another 25% will undergo the more sensitive 3-Tesla MRI, which allows metabolic and physiologic imaging.

A group of 120 patients also will be enrolled in a substudy of Pittsburgh compound B PET scanning. Because the compound binds to amyloid plaques in the brain, it offers a reliable way by which to trace disease progression and may be able to detect early pathologic changes before cognitive changes develop.

At all follow-up visits, patients will donate blood and urine for evaluation for these potential markers; 55% will undergo at least two lumbar punctures for the collection of cerebrospinal fluid. ADNI will focus on four of the most promising biomarkers:

▸ Homocysteine, while not diagnostically significant, may reflect disease progression.

▸ Isoprostanes, indicators of oxidative damage, have been shown to be increased in the hippocampi and CSF of AD patients.

▸ Phosphorylated tau and amyloid-β are the neuropathologic hallmarks of AD tangles and plaques, respectively.

Most research indicates that patients with low levels of amyloid-β and high levels of tau in CSF are more likely to have AD, although some recent studies have challenged this idea.

In fact, none of the biomarkers or imaging modalities included in the ADNI study has been validated in large numbers of patients. Without validation, none can be used as a primary outcome in drug trials, leaving researchers to fall back on imprecise cognitive measures for their main assessment of efficacy, Dr. Trojanowski said.

“If, instead, we could show that biomarkers were changing during a drug trial, in the same way that we can show changes in cholesterol or blood sugar in response to drugs, then we could have a very, very powerful new tool,” Dr. Trojanowski said.

ADNI is an important study in a worldwide effort to discover clinical markers of AD, the investigators agreed. Australia, Japan, and some European countries are undertaking similar projects. Termed “World Wide ADNI,” this informal network will facilitate the performance of international treatment trials and, ultimately, the approval of disease-modifying treatments worldwide. Researchers from these studies are working to harmonize the methodologies, increasing the statistical power of their combined results, Dr. Trojanowski said.

Dr. Weiner noted before ADNI has evaluated even a single patient, it has already changed the way scientists are attacking AD research. Most AD biomarker and imaging research now consists of small studies, each with a unique methodology. Researchers hold their information close to the vest until publication, with months and sometimes years elapsing between the study's conclusion and the sharing of its results.

Not only will all 57 ADNI study centers follow consistent protocols for all the imaging studies and biomarker collection, but also those protocols will be available to any researcher pursuing an independent study. The sharing of methodology and, eventually, of results, is another of ADNI's unique characteristics, Dr. Weiner said. All data will be sent to centralized storage hubs, which will be freely available to anyone–researcher, physician, patient, or family member–who applies for access. The information won't be subject to embargo; scans and biomarker measurements will be available online as soon as they're processed.

“We are going to share everything we get in an unprecedented way. Allowing other researchers to have immediate access to the data is going to maximize this study's effect.”

For more information, including filing a request for access to eventual data, physicians should go to www.adni-info.org

Availability of Markers Bound to Raise Thorny Questions

The first benefits of biologic or imaging markers would be felt in research, where they could hasten the development of disease-modifying drugs. But once those drugs are available, such markers also will be used to identify people most likely to benefit from them and could become part of screening to identify those at risk of developing Alzheimer's disease years before symptoms emerge.

“If and when these therapies become available, who will get them?” asked Dr. Ronald Petersen, director of the Mayo Clinical Alzheimer's Disease Research Center, Rochester, Minn. “Are we going to give them to people with memory impairment first? Or ultimately to those who are asymptomatic but who are at risk? And how will we know who these people are?”

Dr. Petersen hopes eventual findings from his own study help answer some of these questions. He is the lead investigator for the Mayo Clinic Study of Aging, a longitudinal study of 2,000 people aged 70–89 years. The National Institute on Aging provides the study's $1.5 million annual funding package.

In addition to characterizing what brain aging looks like, the study will examine the influence of genetics, family history, and medical comorbidities on the risk of developing Alzheimer's. It will combine this information with cognitive testing, biomarkers, and different imaging modalities in an effort to construct a multivariate model that could predict which apparently healthy people will develop mild cognitive impairment and who might then progress to Alzheimer's.

The Clinical Study of Aging is not related to Alzheimer's Disease Neuroimaging Initiative (ADNI), but both aim to identify the most informative biomarkers and imaging techniques for AD. Eventually, Dr. Petersen predicted, data from both studies might be used to construct a layered screening technique that could identify those who would benefit from early disease prevention or disease-modifying therapy.

Such a model probably would progress from least- to most-invasive testing depending on individual risk, he said. Patients flagged as moderate or high risk might then receive a functional brain scan and a lumbar puncture for cerebrospinal fluid marker sampling. Amyloid imaging with Pittsburgh compound B might also be part of the work-up.

In stratifying patients for treatment, “Earlier may be better than later,” said Dr. John Morris, director of the Alzheimer's Disease Research Center at Washington University, St. Louis. “We know that by the time people start experiencing memory problems, the neuropathological lesions of AD [plaques and tangles] already are present in the brain and, in brain regions vulnerable to AD, nerve cells and their synapses have been lost. Even if truly effective drugs are developed, giving them to people with symptomatic AD–even to those with mild cognitive impairment–may be too late because, by the time symptoms appear, the brain already has been damaged.”

ADNI does not include what could be the critical population in this theory–younger normal controls. “While the study does include nondemented patients (200 normal individuals and 400 with MCI), these people are all older than 55. It is still speculative to consider when the AD process begins in the brain, but it may be in midlife or even earlier.”

To be truly effective in combating the disease, he suggested, it will be important to study younger groups to detect the beginnings of AD with imaging or biological markers before symptoms appear so that eventually therapies hopefully can be given to prevent the occurrence of dementia.

“It's similar to treating atherosclerosis,” he said. “It's a lot easier to treat cholesterol levels to prevent vascular damage than to wait until a heart attack occurs and try to fix things from that point.”

SAN FRANCISCO — Chest pain patients who have a negative CT coronary angiogram can be safely discharged home with an extremely low risk of poor cardiac outcomes at 30 days, Dr. Anne Marie Chang said at the 12th International Conference on Emergency Medicine.

“Prior algorithms have been unable to identify a cohort of patients who have less than a 1% risk of adverse events after discharge,” said Dr. Chang, an emergency medicine resident at the Hospital of the University of Pennsylvania, Philadelphia. “While observation units do allow us to identify this low-risk cohort, they require serial cardiac markers, stress testing, and electrocardiograms, which increase the length of stay in the emergency department.”

Dr. Chang and her colleagues examined 30-day outcomes in a cohort of 568 patients who received CT coronary angiography (CTA) after admission to the ED for chest pain. Half of the patients (285) received their CTA in the ED, with no cardiac markers required; the rest received the procedure after a stay in the observation unit and two or three sets of negative cardiac markers.

All patients had a score of 0–2 on the Thrombolysis in Myocardial Infarction (TIMI) scale. Most patients were black and female; their mean age was 47 years. Hypertension was present in 44% of patients. In two-thirds of patients, the initial electrocardiogram was normal or nonspecific.

Of the 568 patients, 525 had a negative CTA (no lesion of 70% or greater, and a coronary calcium score of less than 100).

In the group of patients who received CTA while in the ED, 81% were discharged home and 19% were admitted, primarily because of confounding medical problems. Stress tests were performed on 13 patients; 2 had positive results and underwent cardiac catheterization. For both of those patients, the catheterization results matched the CTA results; both were medically managed.

In the group that received CTA after an observation unit stay, 97% were discharged home and 3% were admitted. Five patients underwent a stress test, with one positive result. “This patient had a 60% lesion, which was seen on CTA and catheterization, and he received a cardiac stent,” Dr. Chang said at the meeting, which was hosted by the American College of Emergency Physicians.

Most of the cohort with negative CTA results (90%) were contacted by telephone 30 days after discharge. Follow-up on the remainder was performed by a record review and a review of the National Death Index.

The only death resulted from a motor vehicle accident. There were no cardiac deaths, no acute myocardial infarctions, and no revascularizations within the follow-up period.

In addition to accurately identifying those chest pain patients who can be safely discharged from the ED, CTA can also be used as a post-observation-unit discharge strategy, Dr. Chang said.

Studies also show that, compared with stress testing, patients who receive angiography have fewer repeat ED visits and hospitalizations, higher satisfaction with their care, and a better understanding of their disease, Dr. Chang noted.

Patients who receive angiography have fewer repeat ED visits and hospitalizations. DR. CHANG

SAN FRANCISCO — Chest pain patients who have a negative CT coronary angiogram can be safely discharged home with an extremely low risk of poor cardiac outcomes at 30 days, Dr. Anne Marie Chang said at the 12th International Conference on Emergency Medicine.

“Prior algorithms have been unable to identify a cohort of patients who have less than a 1% risk of adverse events after discharge,” said Dr. Chang, an emergency medicine resident at the Hospital of the University of Pennsylvania, Philadelphia. “While observation units do allow us to identify this low-risk cohort, they require serial cardiac markers, stress testing, and electrocardiograms, which increase the length of stay in the emergency department.”

Dr. Chang and her colleagues examined 30-day outcomes in a cohort of 568 patients who received CT coronary angiography (CTA) after admission to the ED for chest pain. Half of the patients (285) received their CTA in the ED, with no cardiac markers required; the rest received the procedure after a stay in the observation unit and two or three sets of negative cardiac markers.

All patients had a score of 0–2 on the Thrombolysis in Myocardial Infarction (TIMI) scale. Most patients were black and female; their mean age was 47 years. Hypertension was present in 44% of patients. In two-thirds of patients, the initial electrocardiogram was normal or nonspecific.

Of the 568 patients, 525 had a negative CTA (no lesion of 70% or greater, and a coronary calcium score of less than 100).

In the group of patients who received CTA while in the ED, 81% were discharged home and 19% were admitted, primarily because of confounding medical problems. Stress tests were performed on 13 patients; 2 had positive results and underwent cardiac catheterization. For both of those patients, the catheterization results matched the CTA results; both were medically managed.

In the group that received CTA after an observation unit stay, 97% were discharged home and 3% were admitted. Five patients underwent a stress test, with one positive result. “This patient had a 60% lesion, which was seen on CTA and catheterization, and he received a cardiac stent,” Dr. Chang said at the meeting, which was hosted by the American College of Emergency Physicians.

Most of the cohort with negative CTA results (90%) were contacted by telephone 30 days after discharge. Follow-up on the remainder was performed by a record review and a review of the National Death Index.

The only death resulted from a motor vehicle accident. There were no cardiac deaths, no acute myocardial infarctions, and no revascularizations within the follow-up period.

In addition to accurately identifying those chest pain patients who can be safely discharged from the ED, CTA can also be used as a post-observation-unit discharge strategy, Dr. Chang said.

Studies also show that, compared with stress testing, patients who receive angiography have fewer repeat ED visits and hospitalizations, higher satisfaction with their care, and a better understanding of their disease, Dr. Chang noted.

Patients who receive angiography have fewer repeat ED visits and hospitalizations. DR. CHANG

SAN FRANCISCO — Chest pain patients who have a negative CT coronary angiogram can be safely discharged home with an extremely low risk of poor cardiac outcomes at 30 days, Dr. Anne Marie Chang said at the 12th International Conference on Emergency Medicine.

“Prior algorithms have been unable to identify a cohort of patients who have less than a 1% risk of adverse events after discharge,” said Dr. Chang, an emergency medicine resident at the Hospital of the University of Pennsylvania, Philadelphia. “While observation units do allow us to identify this low-risk cohort, they require serial cardiac markers, stress testing, and electrocardiograms, which increase the length of stay in the emergency department.”

Dr. Chang and her colleagues examined 30-day outcomes in a cohort of 568 patients who received CT coronary angiography (CTA) after admission to the ED for chest pain. Half of the patients (285) received their CTA in the ED, with no cardiac markers required; the rest received the procedure after a stay in the observation unit and two or three sets of negative cardiac markers.

All patients had a score of 0–2 on the Thrombolysis in Myocardial Infarction (TIMI) scale. Most patients were black and female; their mean age was 47 years. Hypertension was present in 44% of patients. In two-thirds of patients, the initial electrocardiogram was normal or nonspecific.

Of the 568 patients, 525 had a negative CTA (no lesion of 70% or greater, and a coronary calcium score of less than 100).

In the group of patients who received CTA while in the ED, 81% were discharged home and 19% were admitted, primarily because of confounding medical problems. Stress tests were performed on 13 patients; 2 had positive results and underwent cardiac catheterization. For both of those patients, the catheterization results matched the CTA results; both were medically managed.

In the group that received CTA after an observation unit stay, 97% were discharged home and 3% were admitted. Five patients underwent a stress test, with one positive result. “This patient had a 60% lesion, which was seen on CTA and catheterization, and he received a cardiac stent,” Dr. Chang said at the meeting, which was hosted by the American College of Emergency Physicians.

Most of the cohort with negative CTA results (90%) were contacted by telephone 30 days after discharge. Follow-up on the remainder was performed by a record review and a review of the National Death Index.

The only death resulted from a motor vehicle accident. There were no cardiac deaths, no acute myocardial infarctions, and no revascularizations within the follow-up period.

In addition to accurately identifying those chest pain patients who can be safely discharged from the ED, CTA can also be used as a post-observation-unit discharge strategy, Dr. Chang said.

Studies also show that, compared with stress testing, patients who receive angiography have fewer repeat ED visits and hospitalizations, higher satisfaction with their care, and a better understanding of their disease, Dr. Chang noted.

Patients who receive angiography have fewer repeat ED visits and hospitalizations. DR. CHANG

NEW ORLEANS — Rather than unlocking a Pandora's box of nattering e-mails, an electronic patient portal that allows messaging and even access to test results can improve patient satisfaction and decrease patient visits.

“Many physicians think that this type of access is frightening,” Dr. Gretchen P. Purcell said at the annual clinical congress of the American College of Surgeons. “They think they'll be barraged with messages, that patients will misinterpret their test results, and that physicians could even be held legally liable if they don't respond in time to an urgent message.”

But health care providers, who are about 10 years behind the curve in the digital world, need to face up to the facts of the 21st century, said Dr. Purcell of the surgery department at the Children's Hospital at Vanderbilt in Nashville, Tenn. “Patients are demanding the same kind of online access to their medical information as they have for all other aspects of their lives. Those health care institutions that do not have a patient portal now probably will within the next 5 years.”

Patient portals can be designed to suit the needs of different practices and to fulfill various functions. At a minimum, they allow patients to pay bills, schedule or change appointments, and request prescription refills. Other portals are more robust and give patients the ability to review medical records, view test results, and send messages to their health care provider, said Dr. Purcell, who is also with the biomedical informatics department at Vanderbilt Medical Center.

Among the most controversial topics are messaging and the ability to access test results, she said.

“Messaging is probably the function physicians fear the most. Many think it's the equivalent of getting and sending personal e-mail, and this brings up all kinds of worries about security and privacy.”

E-mail and messaging, however, are not the same things. Messages don't go to a personal e-mail account; instead, they go to a dedicated in-box. “This message box is routinely checked by an administrative assistant or nurse—someone who can often answer many of the questions, and who would involve the physician only when necessary—similar to phone call triage.”

There also are concerns that these electronic exchanges aren't part of a patient's documented record. “Some portals can make messaging part of the medical record, and some physicians have found ways to charge for this 'online consultation,'” Dr. Purcell said.

It's important to set clear expectations about response time and emergency issues. Most messaging systems tell patients that they may have to wait 2–3 business days for a personal reply and advise them to call 911 for a medical emergency.

It's not unreasonable to assume that electronic communication could allow patients to bombard offices with questions and requests. Although data are still limited, the studies that are out there suggest just the opposite, Dr. Purcell said.

Two studies published in 2005 indicate that messaging increases patient satisfaction without a corresponding increase in workload. The first study randomized 200 patients to secure messaging or usual care. Only 46% of the patients who were given access sent any message; the average was just 1.5 messages per patient per year. Although messaging didn't reduce the number of telephone calls the office got, the number of office visits in the intervention group did go down (Int. J. Med. Inform. 2005;74:705–10).

The second study randomized 606 patients to a patient communication portal or to a Web site with general health information. Only 31% of the patients given access used the portal. The message box received only one message per day per 250 patients. Again, there was no difference in the number of office telephone calls between the groups, but the patients in the portal group reported better satisfaction with communication and overall care, even if they never used the portal (J. Med. Internet Res. 2005;7:e48).

The same study indicated that secure messaging probably would not overwhelm anyone during working hours, Dr. Purcell said. “Patients tended to use the portal during nonclinic hours—the most convenient time for them—with about 73% of messaging occurring from 5 p.m. until midnight.”

Patients may even be willing to pay for the added convenience of messaging, the authors concluded. Of 341 patients surveyed, 162 (48%) were willing to pay for online correspondence with their physician, with $2 cited as the median payment they thought fair.

Patient access to test results is another area of clinician concern, she said. “Obtaining test results is probably the most commonly desired and most commonly used function of a patient portal, and one that makes physicians very nervous,” Dr. Purcell said.

The MyHealthAtVanderbilt system has three tiers of test results—two can be available to patients online. “Some low-risk, high-value test results, such as cholesterol levels, are available immediately, and some results are available with a delay, such as tests that require interpretation in a specific clinical context,” Dr. Purcell said. “But some results, such as cancer pathology and HIV tests, and others that require intensive patient counseling, are never available though the portal.”

NEW ORLEANS — Rather than unlocking a Pandora's box of nattering e-mails, an electronic patient portal that allows messaging and even access to test results can improve patient satisfaction and decrease patient visits.

“Many physicians think that this type of access is frightening,” Dr. Gretchen P. Purcell said at the annual clinical congress of the American College of Surgeons. “They think they'll be barraged with messages, that patients will misinterpret their test results, and that physicians could even be held legally liable if they don't respond in time to an urgent message.”

But health care providers, who are about 10 years behind the curve in the digital world, need to face up to the facts of the 21st century, said Dr. Purcell of the surgery department at the Children's Hospital at Vanderbilt in Nashville, Tenn. “Patients are demanding the same kind of online access to their medical information as they have for all other aspects of their lives. Those health care institutions that do not have a patient portal now probably will within the next 5 years.”

Patient portals can be designed to suit the needs of different practices and to fulfill various functions. At a minimum, they allow patients to pay bills, schedule or change appointments, and request prescription refills. Other portals are more robust and give patients the ability to review medical records, view test results, and send messages to their health care provider, said Dr. Purcell, who is also with the biomedical informatics department at Vanderbilt Medical Center.

Among the most controversial topics are messaging and the ability to access test results, she said.

“Messaging is probably the function physicians fear the most. Many think it's the equivalent of getting and sending personal e-mail, and this brings up all kinds of worries about security and privacy.”

E-mail and messaging, however, are not the same things. Messages don't go to a personal e-mail account; instead, they go to a dedicated in-box. “This message box is routinely checked by an administrative assistant or nurse—someone who can often answer many of the questions, and who would involve the physician only when necessary—similar to phone call triage.”

There also are concerns that these electronic exchanges aren't part of a patient's documented record. “Some portals can make messaging part of the medical record, and some physicians have found ways to charge for this 'online consultation,'” Dr. Purcell said.

It's important to set clear expectations about response time and emergency issues. Most messaging systems tell patients that they may have to wait 2–3 business days for a personal reply and advise them to call 911 for a medical emergency.

It's not unreasonable to assume that electronic communication could allow patients to bombard offices with questions and requests. Although data are still limited, the studies that are out there suggest just the opposite, Dr. Purcell said.

Two studies published in 2005 indicate that messaging increases patient satisfaction without a corresponding increase in workload. The first study randomized 200 patients to secure messaging or usual care. Only 46% of the patients who were given access sent any message; the average was just 1.5 messages per patient per year. Although messaging didn't reduce the number of telephone calls the office got, the number of office visits in the intervention group did go down (Int. J. Med. Inform. 2005;74:705–10).

The second study randomized 606 patients to a patient communication portal or to a Web site with general health information. Only 31% of the patients given access used the portal. The message box received only one message per day per 250 patients. Again, there was no difference in the number of office telephone calls between the groups, but the patients in the portal group reported better satisfaction with communication and overall care, even if they never used the portal (J. Med. Internet Res. 2005;7:e48).

The same study indicated that secure messaging probably would not overwhelm anyone during working hours, Dr. Purcell said. “Patients tended to use the portal during nonclinic hours—the most convenient time for them—with about 73% of messaging occurring from 5 p.m. until midnight.”

Patients may even be willing to pay for the added convenience of messaging, the authors concluded. Of 341 patients surveyed, 162 (48%) were willing to pay for online correspondence with their physician, with $2 cited as the median payment they thought fair.

Patient access to test results is another area of clinician concern, she said. “Obtaining test results is probably the most commonly desired and most commonly used function of a patient portal, and one that makes physicians very nervous,” Dr. Purcell said.

The MyHealthAtVanderbilt system has three tiers of test results—two can be available to patients online. “Some low-risk, high-value test results, such as cholesterol levels, are available immediately, and some results are available with a delay, such as tests that require interpretation in a specific clinical context,” Dr. Purcell said. “But some results, such as cancer pathology and HIV tests, and others that require intensive patient counseling, are never available though the portal.”

NEW ORLEANS — Rather than unlocking a Pandora's box of nattering e-mails, an electronic patient portal that allows messaging and even access to test results can improve patient satisfaction and decrease patient visits.

“Many physicians think that this type of access is frightening,” Dr. Gretchen P. Purcell said at the annual clinical congress of the American College of Surgeons. “They think they'll be barraged with messages, that patients will misinterpret their test results, and that physicians could even be held legally liable if they don't respond in time to an urgent message.”

But health care providers, who are about 10 years behind the curve in the digital world, need to face up to the facts of the 21st century, said Dr. Purcell of the surgery department at the Children's Hospital at Vanderbilt in Nashville, Tenn. “Patients are demanding the same kind of online access to their medical information as they have for all other aspects of their lives. Those health care institutions that do not have a patient portal now probably will within the next 5 years.”

Patient portals can be designed to suit the needs of different practices and to fulfill various functions. At a minimum, they allow patients to pay bills, schedule or change appointments, and request prescription refills. Other portals are more robust and give patients the ability to review medical records, view test results, and send messages to their health care provider, said Dr. Purcell, who is also with the biomedical informatics department at Vanderbilt Medical Center.

Among the most controversial topics are messaging and the ability to access test results, she said.

“Messaging is probably the function physicians fear the most. Many think it's the equivalent of getting and sending personal e-mail, and this brings up all kinds of worries about security and privacy.”

E-mail and messaging, however, are not the same things. Messages don't go to a personal e-mail account; instead, they go to a dedicated in-box. “This message box is routinely checked by an administrative assistant or nurse—someone who can often answer many of the questions, and who would involve the physician only when necessary—similar to phone call triage.”

There also are concerns that these electronic exchanges aren't part of a patient's documented record. “Some portals can make messaging part of the medical record, and some physicians have found ways to charge for this 'online consultation,'” Dr. Purcell said.

It's important to set clear expectations about response time and emergency issues. Most messaging systems tell patients that they may have to wait 2–3 business days for a personal reply and advise them to call 911 for a medical emergency.

It's not unreasonable to assume that electronic communication could allow patients to bombard offices with questions and requests. Although data are still limited, the studies that are out there suggest just the opposite, Dr. Purcell said.

Two studies published in 2005 indicate that messaging increases patient satisfaction without a corresponding increase in workload. The first study randomized 200 patients to secure messaging or usual care. Only 46% of the patients who were given access sent any message; the average was just 1.5 messages per patient per year. Although messaging didn't reduce the number of telephone calls the office got, the number of office visits in the intervention group did go down (Int. J. Med. Inform. 2005;74:705–10).

The second study randomized 606 patients to a patient communication portal or to a Web site with general health information. Only 31% of the patients given access used the portal. The message box received only one message per day per 250 patients. Again, there was no difference in the number of office telephone calls between the groups, but the patients in the portal group reported better satisfaction with communication and overall care, even if they never used the portal (J. Med. Internet Res. 2005;7:e48).

The same study indicated that secure messaging probably would not overwhelm anyone during working hours, Dr. Purcell said. “Patients tended to use the portal during nonclinic hours—the most convenient time for them—with about 73% of messaging occurring from 5 p.m. until midnight.”

Patients may even be willing to pay for the added convenience of messaging, the authors concluded. Of 341 patients surveyed, 162 (48%) were willing to pay for online correspondence with their physician, with $2 cited as the median payment they thought fair.

Patient access to test results is another area of clinician concern, she said. “Obtaining test results is probably the most commonly desired and most commonly used function of a patient portal, and one that makes physicians very nervous,” Dr. Purcell said.

The MyHealthAtVanderbilt system has three tiers of test results—two can be available to patients online. “Some low-risk, high-value test results, such as cholesterol levels, are available immediately, and some results are available with a delay, such as tests that require interpretation in a specific clinical context,” Dr. Purcell said. “But some results, such as cancer pathology and HIV tests, and others that require intensive patient counseling, are never available though the portal.”