Alzheimer's disease can be fairly accurately diagnosed through measurement of concentrations in cerebrospinal fluid of two proteins associated with the illness, a new primary study and a meta-analysis of 51 other studies has found (JAMA 2003;289:2094-103).
Although Alzheimer's disease is the most common form of dementia, definitive diagnosis relies on neuropathological analysis of the brain at postmortem.
Currently the diagnosis is based on excluding other possibilities and is clinically assumed. Alzheimer's disease can be confused with a number of other dementias, such as Lewy body disease and Pick's disease.
Testing of the cerebrospinal fluid may allow doctors to diagnose the illness earlier and to forestall cognitive decline by beginning treatment sooner.
Researchers led by Dr Trey Sunderland of the National Institute for Mental Health, Bethesda, Maryland, measured the concentrations of two types of protein, b amyloid protein and t protein, in the cerebrospinal fluid of 136 patients with Alzheimer's disease and 72 control patients.
These proteins are implicated in the formation of the b amyloid plaques and neurofibrillary tangles that accumulate in the brains of people with Alzheimer's disease.
In a meta-analysis they also looked at an additional 51 studies—17 looking at b amyloid protein and 34 at t protein—in 3133 patients with the disease and 1481 control patients.
The mean age of the patients in the primary study with Alzheimer's disease was 68 (range 44 to 88) years. Patients were referred by their primary doctor, while the controls responded to advertisements. None of the participants were paid for taking part. Care was taken to ensure that the patients with Alzheimer's disease understood what the study entailed.
Patients were evaluated by medical screening, behavioural observation, magnetic resonance scans, neurocognitive profiling, and lumbar puncture to obtain cerebrospinal fluid.
Within a month of the lumbar puncture all participants were further clinically assessed on the mini-mental status examination, the global deterioration scale, and the clinical dementia rating. Medical and reversible causes of dementia, such as vitamin B-12 deficiency, hypothyroidism, and syphilis, were tested for. Computed tomography was used to screen out patients with multi-infarct dementia.
During the course of the study 36 patients died and underwent autopsy. Alzheimer's disease was confirmed in 31 of these patients. In five suspected cases of Alzheimer's disease the patients were found to have other neuropathological disorders: two patients had Lewy body disease, two had cerebrovascular dementia, and one had thalamic dementia. These patients were excluded from the dataset for the analysis of cerebrospinal fluid.
The final study group comprised 100 live patients with "probable" Alzheimer's disease and 31 patients proved by autopsy to have the disease.
The investigators found that patients with Alzheimer's disease had higher levels of t protein and lower levels of b amyloid protein in their cerebrospinal fluid than the healthy controls. Concentrations of t protein were approximately 2.5 times higher in the group of patients with Alzheimer's disease than in the control patients (mean concentration 587 (SD 365) pg/ml versus 224 (156) pg/ml; P<0.001), whereas concentrations of b amyloid in the patients with Alzheimer's disease were half those in the control group (183 (121) pg/ml versus 491 (245) pg/ml; P<0.001). The cutpoints of 444 pg/ml for b amyloid 195 pg/ml for t protein had a sensitivity of 92% and specificity of 89% for distinguishing patients with Alzheimer's disease from controls, which is comparable with rates for clinical diagnosis.
The meta-analysis confirmed the trend. Concentrations of t protein or b amyloid did not correlate with patients' age, their age at onset of disease, or duration of symptoms. Interestingly, number of years of education was associated with lower levels of t protein in the cerebrospinal fluid, indicating a possible protective effect.
The researchers speculated that concentrations of b amyloid protein were lower in patients with Alzheimer's disease because the protein is deposited in the brain, removing it from circulating in cerebrospinal fluid. Conversely, t protein is thought to be released by dying neurons, so concentrations would increase in the cerebrospinal fluid of Alzheimer's patients.
The researchers estimate that up to 85% of suspected cases of Alzheimer's disease can be definitively diagnosed by analysis of these proteins in cerebrospinal fluid.