A prospective study of cognitive impairment in ALS

Patient selection.

Patients were recruited from the Motor Neuron Diseases Clinic at the University of Western Ontario using a convenience sampling technique. All patients conformed to the diagnosis of clinically definite ALS (El Escorial criteria).6 Inclusion criteria included an age of older than 35 years, English as the primary language, the absence of previous neurologic conditions (e.g., head injury, stroke), and the use of amitriptyline. All ALS patients but one were right handed. All control subjects were right handed. Thirteen patients with ALS (11 men, 2 women), age 40 to 74 years at the time of symptom onset (mean, 54.2 ± 9.6 years) with a mean disease duration of 21.1 ± 10.5 months at initial testing (T1), participated in the study. Eight ALS patients had limb onset of the disease whereas five had bulbar onset. Five spousal control subjects also participated in all aspects of the study except for pulmonary function testing and MRS. When possible, a second test (T2) was performed 6 months after the initial assessment. All testing paradigms were performed on the same day, with the neuropsychological testing completed first to avoid excess fatigue. Pulmonary function testing was performed on the day of assessment for each ALS patient.

Neuropsychological evaluation.

Both the patients and control subjects completed a 2-hour standardized battery of neuropsychological tests at T1 and T2. Tests were chosen to minimize the requirements of speech production and manual motor movement; to be reliable, valid, and sensitive; and to allow for an adequate range of performance. The panel of tests offered breadth in that most major realms of cognition were assessed. Not all patients completed all tests. For example, if motor deficits were apparent, a patient received the oral version and not the written version of the word fluency task. All raw test scores were compared with age- and education-corrected normative data, and were converted into standard z scores with a mean of 0 and an SD of 1. Only descriptive statistics were performed due to the small and uneven sample sizes. Group means less than −1.0 SDs were considered impaired for the purposes of this exploratory study. Similarly, difference scores greater than 1.0 SD were considered meaningful differences.

Tests utilized in the neuropsychological assessment included 1) executive skills: Wisconsin Card Sorting Test (WCST; one-deck version; perseverative responses [WCSTpr], total errors [WCSTterr]), Controlled Oral Word Association (number of words, CFL), and Thurstone Written Word Fluency (number of words [SWords]); 2) concentration/working memory: Consonant Trigrams Test (ConTrig; mean across three trials); 3) memory: Rey Auditory Verbal Learning Test (total across five trials, delayed recall), Recognition Memory Test (RMT; total number of words recalled [RMTWords], total number of faces recalled ]RMTFaces]); 4) visual–perceptual and visual–constructional skill: Motor Free Visual Perception Test (MFVPT; total correct), Block Design Test (Wechsler Adult Intelligence Scale–revised age-scaled score); and 5) mood/personality functioning: Geriatric Depression Scale (GDS; number of items endorsed) and the Neuropsychiatric Inventory (NPI, total score).

Language assessment.

ALS and control subjects completed tests of language and discourse at T1 and T2 (table 17-11). Written responses on the standardized language tests and the discourse tasks were collected and analyzed from subjects whose spoken language was unintelligible.

The standardized language tests comprise a comprehensive set of expressive and receptive measures that reflect linguistic processes thought to be affected by cognitive impairment. Language samples were collected according to the standardized protocol of the “cookie theft” picture description subtest from the Boston Diagnostic Aphasia Examination.9 Discourse samples were collected according to the protocol for topic-directed interviews (TDIs), based on Ripich and Terrell12 in their analyses of the discourse of subjects with AD and other neurogenic disorders.11 The TDIs were initiated using the starter phrase “Tell me about ___________” and included the following five topics: 1) your family, 2) where you were born and raised, 3) your health right now, 4) the jobs you had or the work you did, and 5) what you do each day. Different interviewers collected the TDI data at T1 and T2 to eliminate the effect of shared background knowledge.

All raw test scores were converted to z scores. Student’s t-tests were used to compare the z scores of the language subtest raw scores between ALS patients and control subjects at T1 and T2. Intragroup comparisons also were made between T1 and T2 using paired t-tests. t-Tests also were completed for comparisons of the z scores between bulbar and nonbulbar patients, bulbar patients and control subjects, and limb-onset patients and control subjects. Intragroup comparisons were made for the bulbar and nonbulbar groups between T1 and T2. The alpha level was set at p < 0.10, with familywise corrections made because of the multiple comparisons.

Speech motor control assessment.

Stimulus materials were chosen to describe three levels of speech production—namely, respiratory–laryngeal, resonatory, and oral articulatory. The items included sustained maximum phonation time (MPT), sound repetitions (diadochokinesis [DDK]), and speech intelligibility. These word stimulus items (Word Intelligibility Test [WIT]) were produced once by each subject at each assessment period, with the order and sequence randomized.13 All speech data were tape recorded and analyzed acoustically via special computerized speech analysis software,14 or were analyzed perceptually (WIT) by a panel of listeners. Averages and variability scores were calculated. Initial statistical analyses comparing ALS patients as a group with control subjects were completed employing independent t-tests at the T1 and T2 assessment periods. Subsequently, statistical analyses examined the performance of bulbar- and limb-onset ALS subgroups over time.

MR spectroscopy.

Patients were studied in a Varian (Palo Alto, CA)/Siemens (Erlangen, Germany) Unity Inova 4-T whole-body research scanner. A quadrature radio frequency (RF) coil of our own design was used for transmission and reception.15 The patient’s head was placed in a Plexiglas holder and was packed in firmly using foam to minimize motion. The RF coil was slid over the holder. Global shimming was performed on the whole head, using first- and second-order shims. Proton spectra (¹H) were acquired using a single-voxel short echo time STEAM spectrum (echo time, 20 msec; modulation time, 30 msec; repetition time, 1.5 sec; 256 averages) in each of the anterior cingulate and the right primary motor cortex. The voxel size was 3 mL, and the shim was touched up using the first-order shims in each of the two areas. Water line widths of 8 Hz were typical. Voxels were localized by using a three-dimensional T1-weighted anatomic image, as we did previously.16 Unsuppressed water spectra were also acquired for quantification and spectral referencing. The total study lasted less than 1 hour.

The N-acetylaspartate (NAA) and creatine (Cr) peaks were quantified using Lorentzian line shapes with a peak-fitting algorithm supplied by Varian. Both frequency and line width were allowed to vary. The spectral baseline was flat and required no postprocessing. The reported ratios are ± SE, derived from the variances for each of the three groups, classified by location.

Clinical characteristics.

Thirteen ALS patients (11 men, 2 women) were studied at T1. Eight patients completed the neuropsychology, language, and speech assessments at T2, but only six of these eight patients completed the second MRS study due to sialorrhea or an inability to tolerate recumbency. Three patients died before T2 could be undertaken and one patient withdrew from the study after T1.

The mean age at onset at the time of the first symptom was 54.2 ± 9.6 years (range, 39.9 to 74.0 years). Eleven cases were sporadic and two were familial (in the absence of a known family history of dementia or cognitive impairment). At T1, the mean disease duration from symptom onset was 21.1 ± 10.5 months, with the second assessment occurring after a mean interval of 5.8 ± 0.3 months. ALS patients and control subjects did not differ in age (± 0.28 years, p = 0.78) or years of education (± 0.71 years, p = 0.49). Pulmonary functions demonstrated the patients to be early in the course of their disease state, with no significant impairments observed on measures of forced vital capacity, forced expiratory volume, and peak expiratory flow rates at either T1 or T2. Two of the eight patients could not perform pulmonary function tests at T2 due to physical limitations.

Neuropsychological results.

Language.

Four ALS patients provided written responses on language tests and the discourse tasks at T1. Two patients provided written responses at T2—one who wrote at T1 and one who spoke at T1. The two significant differences in language performance between the ALS and control subjects at T1 and T2 consisted of significantly poorer confrontation naming among the ALS subjects at T1 (t[13] = −3.12, p = 0.008) and poorer single-word vocabulary comprehension (Peabody Picture Vocabulary Test–III [PPVT-III]; t[16] = −2.28, p < 0.05). An analysis of naming errors shows that the ALS patients produced verbal paraphasias (e.g., “yell” for target word “funnel”) and semantic paraphasias (e.g., “nut” for target word “acorn”). ALS patients produced significantly fewer self-corrective utterances than control subjects at T2 on the discourse tasks (t[10] = −2.63, p < 0.05). There were no significant changes in language over time for the following subgroup comparisons: ALS (all patients) versus control subjects, bulbar versus nonbulbar patients, bulbar versus control subjects, and nonbulbar versus control subjects. There were no significant differences over time for the bulbar and the nonbulbar patients.

Speech.

Independent t-tests of MPT revealed that as a group, ALS patients did not differ significantly from the control subjects at T1 (t[15] = −1.14, p = 0.13) and at T2 (t[10] = −0.80, p = 0.28). Paired t-tests used to compare ALS patient MPT at T1 and T2 did not yield significant differences between the two periods (t[(6] = −0.42, p = 0.69). Independent t-tests at T1 and T2 showed no significant speech production differences between the bulbar- and limb-onset groups.

When MPT duration for the production of the sustained vowel /a/ was analyzed between values obtained at T1 and T2, there was a decrease in MPT values for three limb-onset patients and the one bulbar-onset patient. Three limb-onset patients showed an increase in MPT at the 6-month follow-up. However, none of these changes were found to be significant using Wilcoxon’s matched-pairs signed rank test (T = 6, p = 0.05).

DDK was compared between the ALS patients and the control subjects at both T1 and T2 using an independent t-test. The results indicated significantly slower rates of repetition of “puh” (t[12.82] = −4.52, p = 0.001), “tuh” (t[12.03] = −4.20, p = 0.001), “kuh” (t[11.80] = −4.19, p = 0.001), and “puh-tuh-kuh” (t[15] = −2.84, p = 0.013) for the ALS group than the control subjects at T1. Similarly, at T2, ALS patients as a group produced significantly slower DDK rates than control subjects for “puh” (t[10] = −2.38, p = 0.039), “tuh” (t[10] = −2.61, p = 0.026), “kuh” (t[10] = −2.71, p = 0.002), and “puh-tuh-kuh” (t[10] = −3.20, p = 0.010).

No significant differences were found between the bulbar- and the limb-onset groups at T1 on any of the repetition tasks. However, both the bulbar- and the limb-onset groups demonstrated slower DDK rates than the control subjects. Similarly, at T2, both ALS subgroups were lower in articulatory repetition rates than the control subjects.

All patients were judged to be more than 90% intelligible at T1. Wilcoxon’s matched-pairs signed rank test did not yield significant changes in mean percent intelligibility scores over time (T = 6, p = 0.05).

MR spectroscopy.

As illustrated in table 3, at T1 MRS demonstrated a significant neuronal loss in both the anterior cingulate gyrus and the motor strip of the bulbar-onset group of patients as reflected by reductions in the NAA/Cr ratio. Although limb-onset patients failed to demonstrate significant neuronal losses at T1, at T2 neuronal loss was evident in the motor region. No bulbar-onset patients were able to complete T2 MRS due to difficulties with sialorrhea.