Decline in naming is among the most common symptoms in PPA, but the rate of decline is highly variable across individuals.  Leukoaraiosis negatively influences naming recovery in post-stroke aphasia and might influence decline in naming in PPA.

We administered the Boston Naming Test (BNT) at baseline and 1 year later in 27 individuals with PPA .  Two raters evaluated leukoaraoisis on baseline FLAIR sequences, using the Cardiovascular Health Study scale (0-9 scale, using templates for each score). We evaluated effects of leukoaraiosis rating, age, education, and baseline BNT on decline measured by change in BNT score, using linear regression.  Further, we evaluated the effects of leukoaraiosis, age, education, sex, baseline BNT, PPA variant, and any speech-language therapy on the dichotomized outcome of decline in BNT vs no change or improvement, using logistic regression.

Independent raters had excellent reliability in rating scans using the CHS scale (Kappa= 0.75).  Together, leukoaraiosis rating, age, education, and baseline BNT score predicted change in BNT (F (4,22) =5.92; p=0.002).  However, only leukoaraiosis rating predicted change in BNT (t=3.63; p=0.001) independently of the other variables.  When we dichotomized outcome to decline versus no change or improvement, and evaluated the influence of leukoaraiosis rating, age, education, sex, baseline BNT, and treatment (received language therapy or not), decline was predicted by all variables together (chi squared = 20.93; p=0.039; pseudo r² =  0.58). However, only leukoaraosis rating independently predicted decline (OR=5.32; 95%CI 1.14-24.8). For every 1 point increase on the CHS rating scale, there was 5.32 higher chance of showing decline on the BNT over one year.