“These and other data indicate that clinical detection of AD and its distinction from normal aging is more difficult in the oldest old,” Kurt Jellinger, Institute of Clinical Neurobiology, Vienna, Austria, wrote to Alzforum in an e-mail. Jellinger was not involved in the study.

Few studies have taken a hard look at AD in the extreme elderly population, but some suggest that brains differ between people in their sixties and those in their nineties. For one thing, younger AD patients have more plaques and tangles than older ones at autopsy, and pathology can build up in normally aging healthy controls (see ARF related news story on Savva et al., 2009). Further, cross-sectional data suggest that reduced brain volume predicts AD better in 60- to 75-year-olds than in 80- to 90-year-old patients, and that cognitive decline, relative to age-matched controls, in younger AD patients surpasses that in older (see Stricker et al., 2011). Holland delved even further into the interaction of age with AD in a large sample, looking at brain shrinkage and cognitive decline over time, along with cerebrospinal fluid biomarkers (CSF).

The research group analyzed 723 people from the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset who were 65 or older at baseline, including 222 healthy controls, 345 people with MCI, and 156 with early-stage sporadic AD. The Clinical Dementia Rating (CDR) was used to differentiate between groups. Participants came in every six or 12 months for clinical assessment and a magnetic resonance imaging (MRI) scan. About half also gave CSF samples at baseline. Those with AD were followed for up to two years, and healthy controls and those with MCI for up to three. This enabled the team to get baseline measures and, using available follow-up data, calculate annual rates of decline for each person. From this, the scientists derived a spectrum of measurements across ages 65 to 90.

What emerged was a convergence of almost every measure as people approached the oldest ages. For instance, atrophy rate in the whole brain, inferior parietal cortex, and retrosplenial cortex distinguished disease categories quite well in the youngest patients, but as people neared 85 and 90 years old, annual rates of change looked quite similar amongst all three groups. The same held true for two cognitive tests; the ADAS-Cog and the MMSE, but not for the CDR. Baseline values of certain CSF biomarkers also converged at older ages. While concentrations of Aβ42, total tau, and phosphorylated tau (p-tau) diverged among healthy 65-year-olds and those with MCI or AD, levels overlapped in the oldest participants. In particular, CSF p-tau, considered a general marker of neurodegeneration, was almost equal in all of the oldest old, irrespective of AD or MCI diagnosis. In addition, annual conversion rates from MCI to AD fell by half as people aged, dropping from about 16 percent in the youngest to about 8 percent in the oldest.

“It shows that AD proceeds more aggressively in younger patients than among older patients, which might explain why the AD burden [plaques and tangles] is similar among older healthy volunteers and older patients with AD," said Thaís Minett, University of Cambridge, UK.

These findings could be important in the clinic, where cognitive decline is currently the principal way to diagnose AD, said Holland. “In older people showing less deterioration [than younger AD patients], a doctor may be hesitant to diagnose AD,” he told Alzforum—though the patient may have the disease. Older patients may need to be more carefully evaluated to ensure that AD symptoms are not mistaken for simple aging, he said.

These results also have implications for clinical trials. Given the similar rates of decline between very old AD patients and age-matched normal controls, sample sizes would need to be very large to detect therapeutic benefit in this age group. Nevertheless, the oldest old should be included in clinical trials to be sure that they tolerate new medications and that the medicines work in that population, said Claudia Kawas, University of California, Irvine.

Kawas and the authors caution that these results may not apply to the general population. ADNI is a rigorous study and enrolls a selective cohort, she said. Younger, physically able 65-year-olds with aggressive disease are more likely to participate than rapidly declining 90-year-olds. “I’m a little concerned about the bias of the sample,” said Kawas. Further, while conversion to AD seems lower in the oldest people, it could be because MCI in old age is due to mixed pathologies, giving the illusion that these people are more stable, she said.

What explains the slower rates of decline in very old AD patients? One possibility is that they actually developed the disease early, but have always had a slow decline. Alternatively, a faster rate of progression in early disease could precede a slower one later in life. “Both are plausible interpretations of the data,” Holland said. In future studies, he hopes to model the full trajectories of all biomarkers for people at different onset ages to tease out the story further. With the current ADNI data he can start that analysis, though it remains to be seen if it has enough data points to flesh out the entire age spectrum, he said.—Gwyneth Dickey Zakaib.

Reference:
Holland D, Desikan RS, Dale AM, McEvoy LK, for the Alzheimer’s Disease Neuroimaging Initiative. Rates of decline in Alzheimer disease decrease with age. PLoS ONE 2012 Aug 2; 7(8)e42325. Abstract