THE PREDEMENTIA ALZHEIMER'S DISEASE SCALE (PAS)



Determination of cut-off scores of the PAS

In the section selection of the variables of the PAS we described how the items of the PAS were selected. We will now explain how we arrived at the cut-off score for each item.

A. Age
The incidence of AD in the general population increases with age (Ott et al., 1998). The incidence of AD in subjects younger than 60 years is very low in both epidemiological settings (Ott et al., 1998; Rocca et al., 1986) and in clinical settings (Visser, 2000), and therefore subjects younger than 60 are given a score of -1. The risk for AD increases steeply after age 75 in both epidemiological settings (Geerlings et al., 1999; Ott et al., 1997; Rocca et al., 1986) and in clinical settings (Visser, 2000), and for this reason subjects older than 75 years are given a score of 2. Epidemiological data show an increase in incidence of AD after age 70 (Ott et al., 1998) but data from the Maastricht Memory Clinic show an increase in incidence of AD already after age 65 (Visser, 2000). Because the data of the Maastricht Memory Clinic are based on subjects with mild cognitive impairment, we chose to give subjects between 65 and 75 years a score of 1.

B. MMSE
A MMSE score >= 28 is generally considered a normal score (e.g., Golomb et al., 1993) and for this reason a MMSE score >= 28 is given a score of 0. A MMSE score <= 25 is associated with a high risk for dementia in the general population and for this reason these scores are given a score of 2 (Braekhus et al., 1995). The remaining scores (26 and 27) are given a score of 1. Since the MMSE score correlates with age and education, it should be corrected for these variables (Crum et al., 1993; Tombaugh et al., 1992). For simplicity of scoring, we dichotomized the effect of age and made the cut-off scores 1 point lower in subjects older than 75 years (Crum et al., 1993). The effect of education was trichotomized (Crum et al., 1993; Maastricht Aging Study , unpublished data). The cut-off score for subjects with a high level of education ( >= 14 years) was made 1 point higher and the cut-off scores for subjects with a low level of education (<= 8 years) was made 1 point lower than the proposed cut-off scores after correction for age.

C. Functional impairment
Global Deterioration Scale (GDS) (Reisberg et al., 1982) stage 3, the Clinical Dementia Rating scale (CDR) (Hughes et al., 1982) score of 0.5, and the diagnosis of minimal dementia (Roth et al., 1986) are given a score of 2 because they are associated with a high risk of AD (Cooper et al., 1996; de Leon et al., 1993; Devanand et al., 1997; O'Connor et al., 1991; Reisberg et al., 1986; Rubin et al., 1989b). Memory complaints (which is equivalent to GDS stage 2) are less strongly but still significantly associated with AD and is therefore given a score of 1 (Geerlings et al., 1999; Schmand et al., 1996; Schofield et al., 1997).

D. Neuropsychological tests
Impairment is defined as a score below the 10th percentile (or above the 90th percentile for speed related tests) according to age-, sex-, and education-corrected norms. If performance is impaired in one test a score of 1 is given, if performance is impaired in two tests a score of 2 is given, and if performance is not impaired in any test a score of 0 is given. Because memory impairment is the most common symptom in the early stage of AD (Almkvist et al., 1998; Fox et al., 1998; Newman et al., 1994; Nielsen et al., 1999; Small et al., 1997a), a score of -1 is given if memory performance is above average regardless of the score on other tests.

E. Medial temporal lobe atrophy
Medial temporal lobe atrophy can be assessed by either volumetry of the hippocampus (both left and right side) or qualitatively. The PAS rating of volume is based on percentile scores. Atrophy is defined as a volume below the 10th percentile (z-score <= -1.28) and is given a score of 2. If the volume is in the highest tertile (z-score >0.44), a score of -1 is given, if the volume is in the middle tertile (z-score 0.44-0.44) a score of 0 is given, and if the volume is in the lowest tertile but above the 10th percentile (-1.28 < z-score <= -0.44) a score of 1 is given. Qualitative rating should be performed according to the method described by Scheltens et al. (1992) or de Leon et al. (1993a). Both sides should be rated and the highest rating should be used for the PAS scoring. Both authors take a qualitative rating >= 2 as an indicator of atrophy and therefore a rating >= 2 is given a score of 2. A rating of 0 is given a score of 0 and a rating of 1 a score of 1. Since the qualitative rating correlates with age (de Leon et al., 1997; Scheltens et al., 1992) the cut-off scores for subjects older than 75 years are 1 point lower. Thus, in subjects older than 75 years a rating of 0 is given a score of -1, a rating of 1 a score of 0, a score of 2 a rating of 1, and a rating of 3 or higher a score of 2.

F. Apolipoprotein E genotype
Carriers of the apolipoprotein (apoE) e3e4 genotype are given a score of 1. Carriers of the apoE e4e4 genotype are given a score of 2 because the risk of AD is much higher in homozygotes than in heterozygotes (Slooter et al., 1998). Data on whether subjects with the apoE-e2e4 genotype have an increased risk of AD are inconclusive and therefore this genotype is given a score of 0.

References



| Scales | PAS | Score | Instructions | Validation | Table | Variables |