Primary progressive aphasia: A comparative study of progressive nonfluent aphasia and semantic dementia
Primary progressive aphasia (PPA), a degenerative disorder, is often misdiagnosed as Alzheimer's disease. Its subtypes, semantic dementia (SD), and progressive nonfluent aphasia (PNFA), are often difficult to differentiate from each other. Our objective was to highlight the differences in the language profiles of patients with SD and PNFA. To bring out these differences, we report two patients with PPA, one with SD and the other with PNFA. They were administered the Western aphasia battery (WAB) and a semantic battery, which assesses semantic memory. The profiles of language impairment on the WAB indicated that the patient with PNFA had syntactic errors in expressive speech but relatively preserved semantics and comprehension, whereas the patient with SD had preserved syntax but made semantic errors in expressive speech, and had impaired comprehension. There were differences in their performance on the semantic battery too. The patient with SD made relatively less errors on confrontation naming, although on the pointing task he failed to point to those line drawings, which he was unable to name on confrontation. In contrast, the finding of the PNFA patient was the reverse of this. Supplementing conventional neuropsychological tests with formal tests for assessment of language functions is useful in the early diagnosis of PPA. The performance of PPA patients on a detailed assessment of language that includes use of formal tests such as the semantic battery helps to differentiate PNFA from SD.
Keywords: Dementia, language, neuropsychology, primary progressive aphasia, semantic
Frontotemporal lobar degeneration has three major clinical syndromes - the frontal variant of frontotemporal dementia with prominent behavioral changes, and semantic dementia (SD), and progressive nonfluent aphasia (PNFA) [the two forms of primary progressive aphasia (PPA)] with prominent language dysfunction.
In its early stages, the 'naming (language) impairment' of PPA is often mistaken for 'knowing (memory) impairment' and diagnosed as Alzheimer's disease (AD). Likewise, early PNFA is difficult to differentiate from SD., We are aware of only one reported case of PPA from India. Our experience of 4 years in a memory clinic suggests that PPA is under/misdiagnosed as AD.
This report presents the detailed evaluation of a PNFA and a SD patient comparing and contrasting their language profiles and highlights the key diagnostic features.
Case 1 with SD and Case 2 with PNFA were right-handed, native Malayalam speakers, underwent routine clinical, biochemical, neuroimaging, neuropsychological [including Weschler's memory scale, Addenbrooke's cognitive examination (ACE), Reitan's trail making] and language [Western aphasia battery (WAB)] evaluations. Tests were administered in Malayalam. Semantic memory was tested using a battery which was a modification of an earlier version. It included the following tests.
1) Attribute identification: Names of four objects/items are provided as verbal stimuli. One of them differs from the rest on a particular semantic character. The participant is asked to tell the odd-item (e.g., 'car' is the odd-item in an array of furniture names - 'chair, bed, almirah, car'). Almirah is a commonly used word for cupboard in India. In case of incorrect or no response, the names are provided as visual stimuli on a card and the subject asked to point to the correct response. Verbal response attracts full and pointing half credit.
2) Naming to description: A set of semantic features, which taken together, are characteristic of a particular item/object, is read aloud. The participant has to tell the name of the item/object (e.g., for the description, 'a household appliance, which works on electricity and keeps things cold', the correct response is 'fridge', or 'freezer' or 'refrigerator'). In case of incorrect or no response, the correct answer with two semantically related distracters is read aloud from a card (e.g., 'fan', 'fridge,' and 'almirah') and the subject has to choose the appropriate response. Spontaneous response attracts full and forced-choice half credit.
3) Verification of semantic attributes: A statement attributing a semantic feature to an item/object is read aloud. The participant is asked if the statement is correct [(verbally 'yes' or 'no') or nod of their head] (e.g., for the statement 'Birds fly using their wings,' the response is 'yes').
4) Sentence completion: An incomplete sentence (with a blank) is read aloud. A word describing a semantic attribute of the subject in the sentence fits the blank. The participant is asked to tell the most appropriate word to fill-in the blank (e.g., for the sentence 'The crow is ______ in color', the correct response is 'black').
5) Confrontation naming: The subject is shown line drawings of many objects/items, one at a time, and is asked to tell the correct name for each.
6) Picture pointing: Many cards, each with a line drawing of an item/object with no two cards alike, are presented as an array of either 10 or 15 cards at a time. The items/objects in each array are semantically/structurally/functionally related to serve as distracters for the target item. The examiner says aloud the name of the target item/object in the array and asks the participant to point to it.
The composite score on the first four tests is used to assess the attribute awareness.
A 65-year-old man with 11 years of formal education, presented with 1-year history of fluent speech, empty in content and containing incorrect names for objects (e.g., asking for a 'brush' instead of a 'towel,' asking for potatoes when pointing to onions). His memory was relatively preserved though insight was poor. There was no significant medical, psychiatric or family history. Examination was unremarkable.
On the WAB his speech was fluent but lacked content. Auditory comprehension of sequential commands was impaired, repetition was preserved and confrontation naming for real objects better than generative naming [Table - 1]. Reading aloud, matching, and copy writing was preserved but reading comprehension and spontaneous writing impaired (he wrote 'the goat is flying a kite'). On the semantic battery, he was poor on the attribute identification. Naming to description was impaired even with cues (e.g., in response to the stimulus 'it is a household appliance, which works on electricity and keeps things cool,' he repeated the statement and when given the options of 'fan,' 'fridge,' and 'almirah,' he chose 'almirah'). He performed better on verification of attributes and sentence completion. Most errors on confrontation naming were either category co-ordinate errors [saying the name of an item belonging to the same semantic category as the target - e.g., ' hippo ' for ' bear, ' ' pigeon ' for ' parrot' (43%)] , or attribute errors [describing a particular feature or demonstrating the use of the target item instead of naming it - e.g., for ' corn ' he said ' it has small seeds, and you roast it ' (17%)]. On picture pointing he failed on those cards, which he was unable to name on confrontation.
A 79-year-old Physics professor, presented with progressive word finding and naming difficulty of 2-year duration. He had no other cognitive or neurological complaints and remained status quo at 30 months into the illness. He continues to read newspapers and books in both Malayalam and English. Examination was unrevealing except for clumsiness while performing tasks using the right hand.
On the WAB his speech was nonfluent consisting of single word utterances or with few content words [Table - 1]. For, e.g., 'Tree, house car..., Then...' His comprehension was impaired only for the most complex commands. On confrontation he spontaneously named nine objects and with phonemic cues the remaining. His generative naming was poor. He was able to read aloud and perform matching tasks but made errors on complex reading-comprehension task. Copy writing and writing to dictation was essentially normal though spontaneous writing was impaired.
On attribute identification he could not verbally respond but pointed correctly when the names were presented on a card and read aloud. On naming to description he named correctly five out of nine items spontaneously and the remaining on forced choice (e.g., when provided with cues of 'fan,' 'fridge,' and 'almirah' he chose 'fridge' for the statement 'it is a household appliance, which works on electricity and keeps things cool'). His performance was good on verification of attributes but poor on sentence completion. On confrontation naming he got only seven correct making mostly Attempted response errors [says 'that... that...' but couldn't verbalize (35%)]. On picture pointing he pointed correctly to all items, including those he was unable to name.
Neuropsychology and imaging results
Scores on verbal and nonverbal tests [Table - 2] shows that both patients are more impaired on verbal tests (except for the nonverbal WMS-Designs and trail making). MRI brain of both showed asymmetric cortical atrophy, maximally in the left perisylvian temporal region [Figure - 1]. SPECT showed reduced perfusion in the frontal and frontotemporal regions [Figure - 2] as commonly found., Both are independent on activities of daily living (ADL), except for the language handicap.
Progressive nonfluent aphasia begins with nonfluent aphasia, agrammatism, phonemic paraphasias, or anomia. Supportive features are stuttering or oral apraxia, impaired repetition, alexia, agraphia, early preservation of word meaning and late mutism. Semantic dementia is characterized by fluent, empty spontaneous speech, loss of word meaning, semantic paraphasias, preserved matching, single word repetition, and ability to read and write orthographically regular words. The pathology in PNFA involves the left perisylvian region asymmetrically and in SD the anterolateral temporal lobes, usually bilaterally.
Both our patients presented with naming difficulty with greater impairment on verbal than on nonverbal tests. Poor performance on WMS-Designs, digit span, and the trail making was possibly due to comprehension deficits in SD and motor impairment in PNFA, although a subtle frontal dysfunction often reported in the early stages of PPA,, cannot be excluded. These clinical features, in conjunction with the imaging findings, confirm the respective diagnosis of SD and PNFA in our two patients.
On the WAB, spontaneous speech of PNFA patient had dropped function words, and SD patient had meaningless sentences. Comprehension (verbal and reading) was better in PNFA but impaired in SD. Progressive nonfluent aphasia patient could write to dictation while the SD patient could only copy write. The WAB aphasia diagnosis was transcortical motor in PNFA and transcortical sensory in SD, broadly differentiating PNFA from SD. Western aphasia battery could not delineate the nature of semantic memory impairment.
On the semantic battery, although confrontation naming was relatively better in SD, greater impairment on attribute identification, naming to description and pointing demonstrates a greater loss of semantic knowledge that is not only for the names of objects but also for their semantic characteristics. In contrast, although patient with PNFA showed a greater impairment on confrontation naming, a preponderance of attempted response errors, a superior performance on attribute identification and the pointing tasks (pointing correctly to even those line drawings, which he failed to name in both) suggests a relative preservation of semantic knowledge in the presence of an impaired access to the phonological word forms. Preponderance of attribute errors in SD, however suggests, that not all semantic attributes of objects whose names are lost are erased until late stages.
This report illustrates that systematic testing of language and semantic memory and a high index of suspicion can prevent misdiagnosis and aid sub typing of PPA.
[Figure - 1], [Figure - 2][Table - 1], [Table - 2]