Neither Fish Nor Fowl—Dementia With Lewy Bodies Often Missed
Perhaps the biggest, and quintessential, representative of a spectrum neurodegenerative disease is dementia with Lewy bodies (DLB). By some counts, this disease is the second most common form of dementia after Alzheimer disease (AD), with patient estimates for its various forms ranging between one and two million in the U.S. (Aarsland et al., 2008; Weisman and McKeith 2007). All the same, DLB has struggled for recognition and research dollars, being squeezed uncomfortably between its two large neighbors AD and Parkinson disease (PD). “DLB research is an unappreciated field,” said Brit Mollenhauer of the Paracelsus-Elena-Klinik in Kassel, Germany.
DLB is a double whammy of a disease. People with DLB have behavioral and memory problems as in AD and, to a varying extent, also suffer motor symptoms as seen in PD. However, the cognitive symptoms of people with DLB tend to fluctuate frequently, their motor symptoms are milder than in PD, and DLB patients often have vivid visual hallucinations and particular visuospatial deficits. In short, DLB is neither AD nor PD, and yet defining its distinct identity has been a challenge. At the 9th International Conference AD/PD held last March in Prague, ample discussion about DLB resonated from an immediately preceding workshop on this disease and its cousin, Parkinson disease dementia (PDD). Co-organized by Mollenhauer and Richard Dodel, the workshop tried to put DLB more firmly on the research map (see Part 1 of this series).
“Scientists who study DLB think it is a very important disease,” said James Galvin of Washington University, St. Louis. “As a set of independent groups, we and others have worked to increase its face time in the dementia world. We fight a battle because given the limited time and resources funders and reviewers have available to cover related conditions, DLB tends to get the short end of the stick.”
Why is that? Part of the reason is that carving distinct disease categories out of a continuum of symptoms and pathologies is inherently arbitrary. Part of it is that multiple labels being advanced by different investigators for multiple similar variants have not helped the branding. In Prague, several scientists noted that if its awkward name was part of DLB’s identity problem, one solution might be to name it after Kenji Kosaka of Houyuu Hospital in Yokohama, Japan (Kosaka et al., 1980). “Kosaka himself called it a different name, but he really described the clinico-pathological entity that we nowadays diagnose as DLB,” said Michael Schlossmacher of Ottawa University, Canada. “Alzheimer disease, Parkinson disease, Kosaka disease would sound consistent and recognizable to me.”
The Kassel workshop was the latest in a series of small international meetings by a consortium of groups interested in DLB and PDD. In 1995, researchers led by Ian McKeith of Newcastle General Hospital met in Newcastle-upon-Tyne, UK, to hammer out consensus diagnostic criteria (McKeith et al., 1996). This spurred diagnosis in specialty settings and provided a basis for gathering incidence and prevalence data there (Zaccai et al. 2005). But the consensus criteria have not widely penetrated community geriatric, neurology, or primary care settings where many patients are still seen, and both the rate and accuracy of DLB diagnosis remain low. “We are very bad at diagnosing DLB. Up to half of cases diagnosed as DLB turn out at autopsy to have had AD. Misdiagnosis of DLB as AD occurs, as well,” said David Brooks of Imperial College, London.
This has serious consequences. The U.S. Food and Drug Administration, for example, does not formally recognize DLB as a distinct disorder. The agency has a point, Galvin concedes. “They ask: Is it AD? Is it PD? What exactly is it? They ask: Do doctors recognize DLB as separate? No? Then how can you run drug trials, and how can you get doctors to prescribe a future DLB drug?” Galvin said.
With their series of DLB/PDD workshops, McKeith’s and other groups aim to forge a common research agenda that can move their nascent field forward in a concerted way. Subsequent to Newcastle, workshops in the Dutch city of Amsterdam and Yokohama, Japan, continued the effort; and this year’s gathering in Kassel was to lead up to an official centenary workshop in 2012 that will celebrate Frederick Henry Lewey’s first description of Lewy bodies in 1912. (Born in Berlin as Fritz Heinrich Lewy, this German Jewish neurologist-cum-pathologist in his early years worked with Emil Kraepelin and Alois Alzheimer (see Centennial Alzheimer story), but was forced in 1933 emigrate to England and then the U.S.) This year’s workshop designated working groups for biomarkers and for clinical trials, said Mollenhauer. It also included representatives from national DLB societies and patient groups in an effort to help these lay groups beef up their operations such that they can become larger funders of research and lobby for federal funding and recognition, akin to what the Alzheimer’s Association has accomplished for its disease.
Below are some of the main problems, and points of consensus, about DLB from the Kassel and Prague meetings. Part of the reason why more research groups have not taken up focused study of DLB is its complexity. DLB is marked by overlap with AD and PD on the clinical level and on the postmortem pathology level. But clinic and pathology do not match up to a clean picture, leaving the scientist to juggle a welter of descriptive facts that for many fail to “gel” into a tangible entity. Eventually, the solution to this problem will come with new biomarker-driven diagnoses (see Part 5 and Part 6 of this series), but even in the meantime, clinico-pathological correlations have come a long way, the Kassel workshop made clear.
Clinicians agree that people with mixed pathologies suffer faster and more severe disease. Pure Lewy body pathology exists in 10 to 20 percent of cases with clinical DLB, but the majority of patients also have amyloid pathology and many have tangles to varying degrees, as well. Some even have aggregates of TDP43, though whether that is functionally important is not known yet (Arai et al., 2009; Nakashima-Nasuda et al., 2007). It is beyond dispute, however, that mixed pathologies compound each other. Several groups have found that when they looked at postmortem pathology and compared the clinical and cognitive course of the respective patients during their lives, the mixed cases always performed more poorly and progressed faster. Time to nursing home placement, time to death, visuospatial deterioration—whatever the outcome, the mixed cases fared worse. “Pure and mixed clearly are different diseases,” Galvin said.
In the year 2005, the 1996 DLB consensus criteria were revised to focus on the spectrum of Lewy body disorders and to explain more clearly the links between symptoms and pathology. The clinical aspects that set DLB apart from AD, for example, were ascribed to α-synuclein pathology. In a nutshell, this is what the criteria said: DLB and AD share amyloid pathology; people with DLB have α-synuclein pathology, as well, but generally few neurofibrillary tangles. The more tangles a person has, the more their clinical picture overlaps with AD; the fewer tangles they have, the more it diverges from AD.
In the past four years, several groups have further sharpened the cognitive profile of DLB. The goal is to separate DLB not just from AD but also from PDD, where a patient first has Parkinson disease for some years and then develops dementing symptoms. For example, David Salmon of the University of California, San Diego, showed in Kassel that despite some general similarities between DLB and PDD, PDD is marked by deficits in psychomotor speed and attention, which probably arise as α-synuclein pathology spreads from the brainstem via limbic structures and across the cortex. As long as α-synuclein pathology in the cortex remains mild, PD patients tend to stay cognitively intact (Jellinger, 2009). DLB has in common with AD verbal memory deficits driven by these diseases’ shared amyloid and tau pathologies. That is another point of distinction from PDD (Filoteo et al., 2009). But DLB differs from AD by showing pronounced deficits in visuospatial and executive function that Salmon attributes to a unique combination of cortical amyloid and α-synuclein pathology. These differences are useful early on in disease; in late stages the clinical picture of these diseases increasingly merges. Overall, visuospatial tests seemed the most useful for picking out patients with DLB, and these tended to be the people most likely to deteriorate quickly (Hamilton et al., 2008). They also tended to be the ones who suffered visual hallucinations. Beyond these means, disentangling in more detail which clinical features arise from AD pathology and which ones from α-synuclein pathology will require α-synuclein-based and AD pathology-based biomarkers (see also Lippa et al., 2007).
For his part, Galvin and colleagues recently developed cognitive profiles that distinguish AD from healthy brain aging (Johnson et al., 2008). Compared against these profiles, a group of DLB patients performed quite differently from AD, as well. Combining these cognitive data with clinical and amyloid imaging data, Galvin has devised a clinical risk score that detects LBD. “This gives us a good separation. We find that some people who were clinically diagnosed with AD turn out to probably have DLB,” Galvin said. In the laboratory, his group is working on cerebrospinal α-synuclein detection to support this prediction. The goal is to be able eventually to define preclinical DLB. This would work such that a person who is positive for brain amyloid by PET imaging but is cognitively normal could receive CSF biochemistry testing for Aβ/tau and for α-synuclein to determine whether (s)he will likely go on to develop AD or DLB. “We have great confidence in predicting AD based on the PIB/CSF Aβ-tau combination. We want to achieve the same confidence to predict DLB,” Galvin said. For more on that, see Part 6 of this series.—Gabrielle Strobel.
This is Part 3 of a nine-part series. See also Part 1, Part 2, Part 4, Part 5, Part 6, Part 7, Part 8, Part 9.