AD-Diabetes Link Hot Topic at International AD Conference

October 1, 2006

A possible link between diabetes and cognitive dysfunction, specifically Alzheimer disease (AD), is becoming increasingly apparent. Indeed, compared with AD risk in healthy persons, the risk of AD development is 65% to 100% greater in persons with diabetes.

A possible link between diabetes and cognitive dysfunction, specifically Alzheimer disease (AD), is becoming increasingly apparent. Indeed, compared with AD risk in healthy persons, the risk of AD development is 65% to 100% greater in persons with diabetes.1,2 Several studies on the topic, including the use of diabetes drugs in AD management, were reported this summer at the 10th International Conference on Alzheimer Disease and Related Disorders in Madrid, July 15-20.

Researchers from the Karolinska Institute and Stockholm Gerontology Research Center reported that patients with borderline diabetes may be at increased risk for the development of dementia and AD. Weili Xu, a doctoral student, and colleagues from the center monitored 1173 persons aged 75 years and older who did not have dementia or diabetes at baseline. Of these patients, 47 were prediabetic, defined as a casual plasma glucose level between 7.8 and 11 mmol/L. Each patient was examined every 3 years using the DSM-III-R diagnostic criteria.

After 9 years of follow-up, dementia developed in 397 persons; in 307, AD also developed. Of the 47 prediabetic patients, 70% had an increased risk of the comorbid development of dementia and AD. Of interest, a significant association was present only among noncarriers of the apolipoprotein E e4 allele (APOE e4). The risk of AD was especially high when borderline diabetes occurred with severe systolic hypertension.

Similar results were seen in a study by a team led by Rachel Whitmer, PhD, a research scientist at Kaiser Permanente in Oakland, California. This study showed that poor blood sugar control might also be a predictor for cognitive impairment. The researchers observed 22,852 patients listed in the Kaiser Permanente of Northern California Diabetes Registry who had type 2 diabetes for 8 years to determine whether glycosylated hemoglobin (HbA1c) is associated with a risk of incident dementia. The time to dementia diagnosis was analyzed with Cox proportional hazards models adjusted for age, education, race, sex, body mass index, diabetes duration, diabetes treatment, and cardiovascular disease. Compared with persons who had normal HbA1c levels (lower than 7%), patients with levels that were higher than 12% had a 22% higher risk of the development of dementia. Those with levels higher than 15% had a 78% higher risk.

"We know from population studies that patients with diabetes are more likely to have dementia, but this study was unique in that everyone already had diabetes. When diabetes was very poorly controlled, we found that patients were at greater risk for dementia," said Whitmer. "We're not at the point yet where we can say that we can use diabetes drugs to treat cognitive impairment or dementia, but these studies point to other mechanisms, in that research should continue to look at the whole question of diabetes: what's involved with it and why it would effect the brain and cognitive decline."

Michal Schnaider Beeri, PhD, assistant professor of psychiatry, and colleagues at the Mount Sinai School of Medicine in New York City, reported on a study that showed that advanced glycation is associated with cognitive impairment. They evaluated 94 men and 76 women who were cognitively healthy and aged 75 years or older using tests of memory and thinking ability. They also measured advanced glycation end product (AGE) blood levels by obtaining serum low molecular weight carboxymethyllysine (CML). The investigators found that persons with the highest AGE levels did not perform as well as persons with low AGE levels on 10 different tests: Mini-Mental State Examination, immediate recall, delayed recall, Trail Making Test A, Trail Making Test B, Shipley Vocabulary Test, recognition, Boston Naming, fluency, and Digit Symbol. No obvious risk factors were responsible for this relationship other than patient age. On average, women had higher AGE levels than men, and the women with the highest AGE levels had poorer cognitive functioning than women with lower AGE levels. This may account for the substantially higher rate of AD in women compared with men.

In commenting on these studies, Ronald Petersen, MD, PhD, vice chair of the Alzheimer's Association's Scientific Advisory Council told Applied Neurology, "This research is leading us to a point where we can step back and examine which mechanisms link these diseases and, more importantly, whether there are some preventive actions that we can use to reduce the risk of diabetes as well as AD. That would be a win-win situation for the entire population."

THE ROLE OF INSULIN-DEGRADING ENZYME
Three major genes (APP, PSEN-1, PSEN-2) have been implicated in the early-onset familial form of AD, and APOE e4 is a major genetic risk factor for late-onset sporadic AD. Yet, APOE e4 accounts for only about 50% of the genetic variation in late-onset AD.

A number of studies have identified a possible role of insulin-degrading enzyme in the degradation of amyloid-ß protein and the intracellular amyloid precursor protein domain released by g-secretase processing.3 Now, researchers are studying ways to identify and analyze the genes that encode the amyloid-ß processing pathway.

Steven Younkin, MD, PhD, chairman of the Department of Pharmacology, and colleagues at the Mayo Clinic College of Medicine in Jacksonville, Florida, presented findings on the genetics of insulin-degrading enzyme located on chromosome 10q23.3. They are finding that some forms of the enzyme may be associated with an increased or decreased risk of late-onset AD.

The researchers analyzed variants in the conserved region of multiple genes in the amyloid-ß processing pathway, including the gene that encodes the insulin-degrading enzyme, which degrades amyloid-ß as well as insulin. Their analysis demonstrated that the enzyme has 3 protective and 2 risky haplotypes that show replicable association. This study is still under way, and the researchers plan to assemble a large case-control series in the future.

DIABETES DRUGS FOR AD TREATMENT
Diabetes drugs may have a role in the treatment of AD, according to several presenters at the conference. Findings suggest that the 2 thiazolidinediones (TZDs) that have been FDA-approved for type 2 diabetes, pioglitazone (Actos) and rosiglitazone (Avandia), may help reduce new symptoms in patients with cognitive dysfunction.

A team led by Donald Miller, ScD, associate professor of public health at Boston University, reported that AD was nearly 20% less likely to develop in patients who had been using TZDs than in patients who had been using insulin alone. Similar results were found in a separate comparison between TZD users and patients using metformin. The purpose of the study, which looked at 142,328 diabetic patients in a US Department of Veterans Affairs database who received a first prescription for TZD or insulin, was to evaluate whether TZDs could help prevent AD.

"It is still too early for clinical recommendations, but we're getting closer to the point where we can accept the results of these studies with higher confidence," said Miller. "One compelling idea from this research is recognition of new risks for insulin resistance. The resulting high insulin levels may interact with insulin-dependent enzyme, one of a group of enzymes that is involved in clearing amyloid-ß from CNS cells."

Miller and colleagues also examined the rate at which these patients with diabetes were placed in long-term-care facilities. Again, they found lower rates for entering long-term care among patients who were taking TZDs than among patients taking insulin or metformin. "Again, these rates were slightly lower, yet significant," said Miller.

"We are still not sure how these drugs work in this capacity," said David Geldmacher, MD, associate professor of neurology in the Department of Neurology, University of Virginia, Charlottesville. "It will require more preliminary studies to determine what mechanism they are using to deliver their potential benefit."

Geldmacher and his team reported on a clinical trial of pioglitizone for nondiabetic patients with AD to establish the safety of long-term use in this population. Geldmacher said this trial was based on animal studies4 that indicated that pioglitazone and related agents should reduce the brain's inflammatory reaction to amyloid protein deposition.

Geldmacher and colleagues hypothesized that treatment with the drug would not improve function, but it would slow down the progression of AD. Twenty-five patients participated in the study; 12 were treated with 45 mg/d of pioglitazone and 13 received placebo. The drug was well tolerated; the principal adverse effect was limb edema, which is the most common complication associated with pioglitazone in diabetic patients.

Although no statistically significant positive effects on cognition or behavior were seen, Alzheimer's Disease Assessment Scale-Cognitive (ADAS-Cog) scores were an average of 3.5 points better in patients taking pioglitazone compared with placebo at the 1-year point. Geldmacher said the improved mental function was roughly equivalent to that seen with existing therapies, such as cholinesterase inhibitors. Based on the results of this study, Geldmacher and colleagues are planning a 1-year trial of pioglitizone in a larger group of patients with AD.

Treatment with rosiglitazone also may provide similar effects. In speaking with Applied Neurology, Geldmacher cited a recent study by Risner and colleagues5 that tested the effectiveness of rosiglitazone in patients with mild to moderate AD. Patients were randomized to receive 2, 4, or 8 mg of rosiglitazone daily or placebo. After 24 weeks, the 323 patients that completed the study were evaluated. Exploratory analyses demonstrated significant improvement in ADAS-Cog scores in APOE e4-negative patients who were taking 8 mg of rosiglitazone. APOE e4-positive patients did not show improvement and those who received the lowest rosiglitazone dose showed a decline in ADAS-Cog scores. Risner and colleagues proposed that rosiglitazone acts to increase the metabolic efficiency and number of mitochondria.

"The proposed mechanism of action for rosiglitazone is different from what we proposed for pioglitazone at the time that we put the trial together," said Geldmacher. "Rosiglitazone is not known to cross the blood-brain barrier, so its effects on the brain's inflammatory response may not be the way that it's working. It could be working through other mechanisms, such as insulin sensitization in the periphery or reregulation of glucose and insulin metabolism. So, the rosiglitazone results call into question whether our proposed mechanism of action for pioglitazone is actually the mechanism of action that is working for patients with AD. That's the reason for not going out and putting all your patients with AD on pioglitazone."

EARLY INTERVENTION INITIATIVES
The need for effective prevention and treatment is growing because the number of patients with AD will increase within the next few decades as the baby boomers age, said Petersen. The number of new cases of type 2 diabetes among middle-aged Americans has reached epidemic proportions, doubling over the past 30 years,6 and the number of persons with AD is already estimated to exceed 18 million worldwide.7

According to Whitmer, the implications of this research could be of greatest importance for primary care physicians and endocrinologists to keep in mind. "Endocrinologists need to understand that diabetes can cause cognitive impairment, and that's been a new concept for them to get their arms around," she told Applied Neurology. "The problem with the neurologists is that they see people when they're farther down the pathway. They're not seeing people when they're middle-aged, overweight, and prediabetic."

Neurologists should recommend aggressive management of diabetes to patients with AD, said Geldmacher. "This does not mean that neurologists should get into the habit of recommending specific diabetes drugs for the treatment of cognitive dysfunction," he said. "When talking to my patients, I tell them it's important to follow through with a healthy diet and exercise."

For more information, physicians can also refer their patients to the Alzheimer's Association, which is distributing patient information on this topic on their Web site, through a number of presentations, and in its recent Maintain Your Brain booklet that emphasizes the importance of a healthy diet and exercise.

REFERENCES1. Ott A, Stolk RP, van Harskamp F, et al. Diabetes mellitus and the risk of dementia: the Rotterdam Study. Neurology. 1999;53:1907-1909.
2. Arvanitakis Z, Wilson RS, Bienias JL, et al. "Diabetes mellitus and risk of Alzheimer disease and decline in cognitive function." Arch Neurol. 2004;61:661-666.
3. Bjork BF, Katzov H, Kehoe P, et al. Positive association between risk for late-onset Alzheimer disease and genetic variation in IDE. Neurobiol Aging. 2006;[Epub ahead of print].
4. Schutz B, Reimann J, Dumitrescu-Ozimek L, et al. The oral antidiabetic pioglitazone protects from neurodegeneration and amyotrophic lateral sclerosis-like symptoms in superoxide dismutase-G93A transgenic mice. J Neurosci. 2005;25:7805-7812.
5. Risner ME, Saunders AM, Altman JF, et al. Rosiglitazone in Alzheimer's Disease Study Group. Efficacy of rosiglitazone in a genetically defined population with mild-to-moderate Alzheimer's disease. Pharmacogenomics J. 2006;6:246-254.
6. Fox CS, Pencina MJ, Meigs JB, et al. Trends in the incidence of type 2 diabetes mellitus from the 1970s to the 1990s: the Framingham Heart Study. Circulation. 2006;113:2914-2918.
7. Alzheimer's Society. Policy Positions. Available at: www.alzheimers.org.uk/News_and_campaigns/Policy_Watch/demography.htm.
Accessed September 6, 2006.