British Journal of Psychiatry (1992), 161, 522—531 The human immunodeficiency virus-i (HIV) is neurotrophic and enters the central nervous system soon after initial infection. One long-term con sequence of this brain penetration is the development of progressive impairment of cognitive and motor function due to a direct effect of the virus on neurons. This was originally called the AIDS dementia complex (ADC; Navia et al, 1986). The relationship between early viral access and the subsequent development of the ADC has provoked controversy because of the description of abnormal neuropsychological test results in HI V-positive but otherwise asymptomatic individuals. A small, but influential, study showed cognitive abnormalities in seven out of 13 asymptomatic HI V-positive patients (Grant et al, 1987). The inunediate inference was that the presence of minor test abnormalities in such patients meant that they were developing dementia. This has led to much larger and perhaps better controlledstudiesof neuropsychological function among HIV-positive patients (Miller et al, 1990; Selnes et al, 1990). These larger studies have prompted the conclusion that the ADC does not have an early or insidious onset (Harter, 1989). Although thismay be trueforthemajorityofpatients with HIV in the early stages of the disease, there have continued to be reports of test abnormalities in small, but well controlled, studies of asymptomatic patients (Lunnetal, 1991; Perry etal, 1989; Stern etal, 1991; Wilkie et al, 1990) and there is little doubt that there is cognitive impairment in later symptomatic stages of the disease(Janssenet al, 1989; Trosset al, 1988; Miller et al, 1990). However, even in symptomatic disease, only a minority of patients will develop clinically significant dementia. The prognosis forsuch patients, and the relationship of neuropsychological function to depressive symptoms in this group (Saykin et a!, 1988; Miller et a!, 1990), remain of practical importance. The main weakness of the existing studies is that they have almost exclusively involved cohorts of homosexual men. Results in this population may not generalise to others. In particular, a different natural history for HIV-related cognitive impairment may be anticipated for drug users. Drug users are one of the largest high-risk groups for infection with HIV and a variety of factors may increase their vulnerability to the effects of infection. Firstly, opiates appear to have immunosuppressive actions that accelerate the immunodeficiency related to HIV (Weber et al, 1990). Secondly, injection drug users are usually of below average inteffigence, poorly educated, frequently give a history of head injury or lossof consciousness relatedto theuseof drugs, 522 The Edinburgh Cohort of HIV-Positive Drug Users: Pattern of Cognitive Impairment in Relation to Progression of Disease VINCENT EGAN, RAY P. BRETTLE and GUY M. GOODWIN To examine the neuropsychiatric effects of infection with HIV,220 drug users (27 HIVnegative, 193 HIV positive) completed tests evaluating premorbid intelligence, memory, non-verbal performance, information processing speed, and mood. When these measures were compared cross-sectionally by the severity of HIVillness, symptomatic patients (in CDC stage IV)were impaired on Trails B, two-choice decision time, delayed recall of the Wechsler Logical Memory Test and most components of the AuditoryVerbalLearningTest. These findings implyreduced capacity for concentration, speed of thought and memory. When 101 patients were retested a mean of 16 months after their initial assessment, performance on Trails A and B, Block Design and delayed recall of the Wechsler Logical Memory Test deteriorated more for patients at, or progressing within, CDCstage IV,than performance of patients at stage Ill.The results broadly correspond to the cross-sectional findings. However, there was a decline in all tests of memory function for the sample independent of clinical staging. This may be evidence of brain involvement before the appearance of other symptoms. Self-rated measures of mood did not change cross-sectionally, progressively, or interactively with time and stage of HIV illness, and cannot account for the changes in cognitive function observed. Change in drug use, similarly, does not account for the cognitive findings. Four (5%) of the retested subjects developed AIDS dementia complex, but most of the performance and memory impairments seen were subclinical despite the destructive neuropathology presumed to underlie intellectual decline in patients with HIV infection. An exploratory analysis of treatment with zidovudine in65 patients withstageIVdiseaseshowed nodemonstrable benefit forcognitive function.