To the Editor:
The identification of HIV-related cognitive impairment remains important given the increased mortality rates (1) and vocational impairments (2) seen among HIV-seropositive individuals presenting with even subtle neurocognitive deficits.
A number of brief screening measures, including the HIV Dementia Scale (HDS) (3), Mental Alteration Test (4), Executive Interview (5,6), and HIV Dementia Assessment (7), have been developed or adapted to screen for HIV-associated dementia. Of these, the HDS has received widespread attention (8-12). Consisting of four components designed to target verbal memory recall, psychomotor speed, visual construction, and response inhibition, HDS scores range from 0 to 16, with scores of 10 or less being regarded as the cutoff for HIV-associated dementia (3). Importantly, in comparison with the Mini Mental State Examination (13) and the Grooved Pegboard (14), the HDS has demonstrated some clinical utility in the classification of HIV-associated dementia, because sensitivity and specificity rates have been reported to range from 80% to 91%, respectively.
The current study sought to validate the utility of the HDS in the identification of subtle HIV-associated cognitive deficits. A recent abstract (15) suggested that performance on the HDS might have some utility in screening for HIV-related cognitive impairment. In contrast to Childers et al. (15), who employed clinical ratings for classification purposes, we used published normative data for classification purposes. We hypothesized that the previously validated HDS cutoff score for the identification of HIV-associated dementia would contribute to an unacceptably high false-negative classification rate.
Ninety HIV-1-seropositive men and women participated in this study. All participants were enrolled in a longitudinal study designed to identify the neuropsychiatric barriers and facilitators among HIV-infected individuals seeking to return to work. All participants planned on returning to work because they did not perceive themselves as being significantly impaired cognitively.
Persons were considered eligible for enrollment in the study program if they met the following inclusion criteria: HIV-1 seropositive; age 18 through 55 years; history of employment for at least a 12-month period in a paid full-time job; stopped work within the context of HIV-related illness; not currently engaged in paid employment; no substance use disorder within the past 12 months; negative toxicology screen for substances such as heroin, cocaine, and nonmedical benzodiazepines; no current or lifetime diagnosis of a severe and persistent mental illness; and not diagnosed with moderate or severe dementia syndrome (e.g., HIV-associated dementia moderate or severe, posttraumatic dementia) using the DSM-IV or American Academy of Neurology (AAN) criteria.
The HDS consists of four subtests: timed written alphabet, recall of four items after a 5-minute delay, timed cube copy, and antisaccadic errors. Because the HDS was designed to be administered by a broad array of clinicians, technical aspects of the administration and scoring were kept to a minimum in the initial publication (3). For the current study, we attempted to standardize the administration and scoring of the HDS by the following methods. First, the administration of the 20 antisaccadic trials was standardized by creating a specific pattern of trials to ensure that 10 trials were administered to both the right and left eyes. Second, we adopted standardized scoring of the cube drawing using scoring criteria from the Developmental Test of Visual-Motor Integration (16), which included the presence of 12 distinct lines and three-dimensional representation.
In addition to the HDS, all participants were administered a screening battery consisting of the Beck Depression Inventory (BDI) (17) and the following seven neuropsychologic (NP) tests: Letter-Number Sequencing subtest from the Wechsler Adult Intelligence Scale-III (18), Stroop Color-Word Test-Golden version (19), Color Trails 1 and 2 (20), California Verbal Learning Test (21), Faces I and Faces II from the Wechsler Memory Scale-III (22), California Computerized Assessment Program (23), and Grooved Pegboard (14).
To evaluate the performance characteristics of the HDS, all subjects were classified as neuropsychologically abnormal or normal based on their respective performance on the NP measures independent of the results from the HDS. Cognitive abnormality was defined as performance that deviated least 2 SD units below established norms on at least two independent NP measures (e.g., Grooved Pegboard and California Verbal Learning Test).
Using these criteria, 46 individuals were classified as being neuropsychologically normal and 44 individuals were classified as being neuropsychologically abnormal. Those classified as neuropsychologically normal had a mean age of 40.9 (±5.4) years and an education level of 13.8 (±2.7) years. The neuropsychologically abnormal group had a mean age of 40.8 (±7.0) years and an education level of 13.1 (±2.3) years. No age, education level, or gender differences were noted between the two groups. Additionally, the two groups did not differ in terms of level of depressive symptoms as reported on the BDI, most recent CD4 count, HIV RNA viral load, or years since HIV-1 seropositivity. Among the entire sample, HDS performance was not related to age, level of education, number of years since known HIV-1 infection, or BDI score.
To evaluate the utility of the HDS to classify individuals identified as being neuropsychologically abnormal, we examined the classification rate of the HDS using the cutoff score of 10, which had previously been shown to classify HIV-related dementia accurately (3). Using this cutoff score, 17 of the 44 of the individuals classified as NP impaired were accurately identified (sensitivity = 39%). In contrast, the HDS performance accurately classified 39 of 46 of the NP normal individuals (specificity = 85%).
The results indicate that in contrast to previously published findings, the HDS lacks sufficient sensitivity to screen for NP abnormality beyond frank dementia. Intact performance (i.e., performance above established cutoff levels) contributes to a significant number of false-negative errors, suggesting that a more complete NP battery should be administered in those cases in which subtle neurocognitive deficits are suspected.
The sensitivity and specificity of the HDS reported in this study are significantly different than previously reported classification rates. In the current study, use of a score of 10 or below on the HDS led to a significant number of false-negative classifications. We anticipated this relationship, because the previous studies evaluated the clinical utility of the HDS in classifying demented versus nondemented HIV-infected individuals. The current study sought to evaluate the utility of the HDS in identifying individuals who present with NP abnormality rather than with frank dementia by excluding those HIV-seropositive individuals diagnosed with dementia. Based on the results of the current study, given that the HDS demonstrated poor sensitivity in identifying neuropsychologically abnormal performance, the utility of the HDS to serve as a screening instrument beyond the identification of frank dementia is limited.
The classification rates reported in the current study are strikingly different from a recent abstract (15), which also evaluated the utility of the HDS in classifying NP abnormality. Employing global impairment ratings, Childers and colleagues (15) reported that the performance on the HDS classified 97% of those deemed neuropsychologically impaired, although a significant number of false-positive errors were reported. The reason(s) for the vastly different classification rates is unclear, although two design issues merit consideration. First, the different classification rates may reflect differing criteria for assigning NP impairment. For example, Childers et al. (15) employed the use of demographically corrected NP test scores for assigning a global impairment rating based on performance in identified cognitive domains, whereas the current study employed a common rule-based quantitative approach to classification by defining impairment as performance at least 2 SD units below published norms on two or more measures. Second, by specifically excluding the most severely cognitively impaired individuals from the current study, it is likely that we reduced the overall sensitivity of the HDS by restricting the range of impairment.
Overall, there remains a clinical need for reliable and brief screening measures to detect and monitor HIV-related cognitive dysfunction as well as to quantify any response to therapeutic interventions. The current study found that the utility of the HDS in detecting HIV-related cognitive abnormalities beyond frank dementia is limited. The current study suggests that using the accepted HDS cutoff score of 10 or less is likely to contribute to a significant number of false-negative errors, which is problematic, given the association of subtle cognitive deficits with increased mortality and vocational instability. Although this does not rule out the utility of the HDS for its designed purpose, the need for the carefully designed NP evaluation remains clinically relevant.
Acknowledgment:
This project was supported in part by National Institutes of Health grants 5T32MH19998-5 and 5R01MH060560-05 (Principal Investigator: W.G.V.G.).
*Clifford A. Smith
†Wilfred G. van Gorp
‡Elizabeth R. Ryan
§Stephen J. Ferrando
†Judith Rabkin
REFERENCES
1. Ellis RJ, Deutsch R, Heaton RK, et al. Neurocognitive impairment is an independent risk factor for death in HIV infection: San Diego HIV Neurobehavioral Research Center Group. Arch Neurol. 1997; 54:416-424.
2. Heaton RK, Marcotte TD, White DA, et al. Nature and vocational significance of neuropsychological impairment associated with HIV infection. Clin Neuropsychol. 1996; 10:1-14.
3. Power C, Selnes OA, Grim JA, et al. HIV Dementia Scale: a rapid screening test. J Acquir Immune Defic Hum Retrovirol. 1995; 8:273-278.
4. Jones BN, Teng EL, Folstein MF, et al. A new bedside test of cognition for patients with HIV infection. Ann Intern Med. 1993; 119:1001-1004.
5. Royall DR, Mahurin RK, Gray KF. Bedside assessment of executive cognitive impairment: the Executive Interview. J Am Geriatr Soc. 1992; 40:1221-1226.
6. Berghuis JP, Uldall KK, Lalonde B. Validity of two scales in identifying HIV-associated dementia. J Acquir Immune Defic Syndr. 1999; 21:134-140.
7. Grassi MP, Perin C, Borella M, et al. Assessment of cognitive function in asymptomatic HIV-positive subjects. Eur Neurol. 1999; 42:225-229.
8. Dougherty RH, Skolasky RL, McArthur JC. Progression of HIV-associated dementia treated with HAART. AIDS Reader. 2002; 12:69-74.
9. Chang L, Ernst T, Leonido-Yee M, et al. Highly active antiretroviral therapy reverses brain metabolite abnormalities in mild HIV dementia. Neurology. 1999; 53:782-789.
10. Chang L, Ernst T, Leonido-Yee M, et al. Cerebral metabolite abnormalities correlate with clinical severity of HIV-1 cognitive motor complex. Neurology. 1999; 52:100-108.
11. Chang L, Ernst T, Leonido-Yee M, et al. Perfusion MRI detects rCBF abnormalities in early stages of HIV-cognitive motor complex. Neurology. 2000; 54:389-396.
12. Von Giesen HJ, Wittsack HJ, Wenserski F, et al. Basal ganglia metabolite abnormalities in minor motor disorders associated with Human Immunodeficiency Virus Type 1. Arch Neurol. 2001; 58:1281-1286.
13. Folstein MF, Folstein SE, McHugh PR. Mini Mental State: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975; 12:189-198.
14. Matthews CG, Kløve H. Instruction manual for the adult neuropsychology test battery. Madison, WI: University of Wisconsin Medical School, 1964.
15. Childers M, Ellis M, Deutsch R, et al. The utility and limitations of the HIV Dementia Scale. J Int Neuropsychol Soc. 2002; 8:160.
16. Berry KE, Buktenica NA. Developmental test of visual motor integration. Odessa, FL: Psychological Assessment Resources, 1997.
17. Beck AT, Beamesderfer A. Assessment of depression: the depression inventory. Mod Probl Pharmacopsychiatry. 1974; 7:151-169.
18. Wechsler D. Wechsler adult intelligence scale-III. New York: The Psychological Corporation, 1997.
19. Golden JC. Stroop color and word test. Chicago: Stoetling Company, 1978.
20. D'Elia LF, Satz P, Uchiyama CL, et al. Color trails test. Odessa, FL: Psychological Assessment Resources, 1996.
21. Delis DC, Kramer JH, Kaplan E, et al. California verbal learning test. San Antonio: The Psychological Corporation, 2000.
22. Wechsler D. Wechsler memory scale-III. New York: The Psychological Corporation, 1997.
23. Miller EN, Satz P, Visscher B. Computerized and conventional neuropsychological assessment of HIV-1-infected homosexual men. Neurology. 1991; 41:1608-1616.
© 2003 Lippincott Williams & Wilkins, Inc.