Their work not only strengthens the idea that AD is a heterogeneous disease, but also raises the specter of faulty drug data if people with rapidly progressing AD enroll in clinical trials, noted Joy Snider of Washington University School of Medicine in St. Louis, Missouri. Snider suggested that rapid progressors could potentially skew trial data, making a compound appear ineffective when, in fact, it might have worked on a much slower time scale than rpAD. Scientists say the time is ripe to begin parsing out what is causing unusually quick decline in this subset of AD patients.

The possibility that some people with fast-progressing dementia might have Alzheimer’s dawned on Schmidt and Zerr when autopsies showed that local patients with suspected prion disease had enough AD pathology to qualify them for a posthumous diagnosis of that disease. Checking medical records on 32 such cases, the researchers discovered they had had symptoms resembling Creutzfeldt-Jakob disease. They dropped five to six points per year on the Mini-Mental State Examination—twice as fast as do typical AD patients. Strikingly, these rpAD patients died about two years after diagnosis, whereas people with AD typically survive about eight years after symptoms emerge (Goldberg, 2007). Schmidt and Zerr published these data in 2010 (Schmidt et al., 2010).

To determine if rpAD is widespread, the scientists then searched the literature for similar cases (see ARF related news story on Schmidt et al., 2011). They expanded their retrospective analysis to include 89 more people who turned out to have neuropathological hallmarks of AD at autopsy after being treated at prion disease surveillance centers in France, Germany, Japan, and Spain (Schmidt et al., 2012). Since the German surveillance center is set up to track CJD but not rpAD, Zerr does not know precisely how many total cases of suspected prion disease were analyzed in order to come up with the 89 rpAD cases in the 2012 analysis. She estimates that people with rpAD make up less than 5 percent of CJD cases.

The rpAD patients did not have other causes of rapid dementia—such as tumors, stroke, inflammation, or extensive vascular disease—and lacked a family history suggesting autosomal-dominant mutations. Reviewing medical charts, the researchers noted that these rpAD patients, on average, developed disease around 74 years of age and died just 10 months after disease onset—more than twice as quickly as the local cohort in the 2010 study.

The discrepant survival figures in the two rpAD studies might be due to difficulties pinpointing disease onset, especially with different protocols across multiple centers, Schmidt and Zerr suggested. As in every study focusing on dementia, the exact determination of the clinical onset—and consequently, the calculation of decline and survival time—is prone to poor memory, the authors noted. “Moreover, the approaches to data acquisition certainly differed among the surveillance centers, which made data pooling and comparison more difficult.” The patients’ unusually fast deterioration may have kept them from making follow-up visits, further complicating matters. “It is possible that rapid decliners are underestimated in AD longitudinal studies,” suggested Walter Schulz-Schaeffer, who works with the authors at Georg-August University. Other researchers contacted by Alzforum are unsure if rpAD truly exists, and say more work is needed to be certain.

What distinguishes rpAD from less aggressive forms of the disease? ApoE4 is a major genetic risk factor for AD, but curiously, the pooled rpAD group had no ApoE4 homozygotes among them. Only 20 percent carried one ApoE4 allele; typically, 40 to 65 percent of AD patients do. The findings suggest ApoE does not contribute to the precipitous speed of this form of AD. Other studies have also correlated lack of ApoE4 with rapid decline. For example, people with familial, early-onset AD deteriorate faster than patients with late-onset AD, and this effect is magnified in ApoE4 non-carriers (Van der Vlies et al., 2009). Annemieke Rozemuller of VU University Medical Center, Amsterdam, the Netherlands, thinks the apparent underrepresentation of ApoE4 could stem simply from the selection criteria used in the retrospective analysis, not be truly meaningful to this form of AD. The study excluded patients with severe vascular disease, yet prior research has linked ApoE4 with cerebral amyloid angiopathy (Greenberg et al., 1995; Premkumar et al., 1996).

Researchers have looked to biomarkers in the cerebrospinal fluid (CSF) to try to explain the rapid progression. Curiously, in their 2012 study, Schmidt and Zerr found that CSF Aβ42, Aβ40, tau, and phospho-tau levels in rpAD patients fell within the range for classic AD. “I would have expected tau and phospho-tau levels to be much higher,” said Henrik Zetterberg of the University of Gothenburg in Sweden. “That they are not speaks against AD being the cause of the rapid degeneration [in these patients],” he said, noting other studies that link high CSF tau with rapid disease progression in people with AD (Kester et al., 2009; Wallin et al., 2010; Sämgård et al., 2010) or mild cognitive impairment (van Rossum et al., 2012; Blom et al., 2009). Interestingly, 42 percent of rpAD patients in the 2012 multicenter study tested positive for CSF 14-3-3, a protein marker for neuronal destruction typically seen in CJD but not in AD. "The 14-3-3 profile indicates the brain is suffering," said Zetterberg.

Cognitive test scores across longer time periods might yield other clues to what drives the sharp downturn. “If we see a six-point decline on MMSE in one year, we don’t know if this is due to a single acute event such as an infarct or cardiac event, or whether something is acting over a longer period to force a short disease course—for example, white matter degeneration due to vascular pathology,” Schulz-Schaeffer said.

Other scientists said more detailed neuropathologic analysis would provide insight into the accelerated disease course of rpAD patients. This is important because there are several distinct clinicopathologic subtypes of AD,” noted Kurt Jellinger of the Institute of Clinical Neurobiology, Vienna, Austria. A recent retrospective analysis found that people with neurofibrillary tangles primarily outside the hippocampus tend to progress faster and die earlier than AD patients with heavy tau pathology in limbic areas (Murray et al., 2011; see also Jellinger, 2012). Dietmar Thal of the University of Ulm, Germany, agrees, stressing the need for “meticulous investigation of the entire brain to find out which types of neurons are affected.” Along those lines, Rozemuller and colleagues have data indicating stronger parietal cortex involvement in young familial AD patients with slowly progressing disease, and more temporal pathology in those who decline faster. She reported these findings at the 2011 Alzheimer’s Association International Conference in Paris, France.

Michael Geschwind of the University of California, San Francisco, suggested looking at ultrastructural properties of the proteinopathy in rpAD patients, as work from the prion field has correlated small fibril size with rapid disease onset (Legname et al., 2006).

Zerr said her team will collaborate with Heiko Braak of the University of Ulm, Germany, to get a closer look at the Aβ and tau pathology in rpAD patients. In addition, she and Schmidt have initiated a prospective study comparing biomarker and genetic profiles in people with classic and rpAD. So far, the scientists have recruited 150 to 200 patients throughout Germany who initially reported with rapid dementia and CJD symptoms but were subsequently diagnosed with AD using the Dubois criteria. Study participants will be genotyped for ApoE, prion protein, and 12 other AD risk alleles. They will come for yearly measurements of brain imaging and CSF biomarkers, as well as cognitive and functional batteries. In addition, the scientists will collect data on risk factors and other demographic and lifestyle information. Eventually, they hope to do GWAS in rpAD, though this would require “at least several hundred patients for meaningful analysis,” Zerr said.—Esther Landhuis.