icon-    folder.gif   Conference Reports for NATAP  
  XIX International AIDS Conference
July 22-27, 2012
Washington, DC
Back grey_arrow_rt.gif
Diabetes Linked to Poor Cognitive Function in HIV-Positive Adults
  XIX International AIDS Conference, July 22-27, 2012, Washington, DC

Mark Mascolini

Diabetes in HIV-positive members of the French Aquitaine cohort was independently associated with lower scores on tests of cognitive function, according to results of a 400-person analysis [1]. No other cardiovascular-related factors affected risk of poor cognitive performance in this cohort.

Plentiful research documents higher rates of cardiovascular morbidity and mortality in HIV-positive people than in the general population. Cognitive function also declines more rapidly in people with than without HIV. To further the study of mechanistic links among these conditions, ANRS-CO3 Aquitaine Cohort investigators conducted this analysis of HIV-positive adults in southwestern France.

During the 2007-2009 study period, the 400 study participants underwent assessment of seven cognitive domains, diabetes, prediabetes, and other cardiovascular variables. Follow-up continued for 2 years in 288 of the 400 participants (72%). Study participants could not have an opportunistic infection or cancer being treated at the time of enrollment.

The Aquitaine team analyzed cognitive performance in seven domains and used the latest definition of HIV-associated neurocognitive disorder (HAND). The researchers considered both raw cognitive test scores and annualized percentage change in cognition. They defined diabetes as blood sugar above 7 mmol/L at least twice, or above 11.1 mmol/L once, or use of antidiabetic drugs. They defined impaired glycemia as at least two blood sugars between 6.1 and 7 mmol/L before enrollment.

Of the 400 cohort members, 320 (80%) were men and 96 (24%) had AIDS. Age averaged 47.3 years, median CD4 count stood at 515 and median CD4 nadir at 260. Most study participants (85%) had a viral load below 500 copies, and 89% were taking antiretrovirals. Thirty-eight people (9.5%) had diabetes, 32 (8%) had prediabetes, 177 (44%) had hypercholesterolemia, and 81 (20%) had hypertension.

Eighty-four cohort members (21%) had asymptomatic neurocognitive impairment, 126 (31.5%) had mild neurocognitive disorder, and 27 (6.75%) had HIV-associated dementia. Rates of normal glycemia, prediabetes, and diabetes were 4.2%, 9.5%, and 11.8% for people with asymptomatic impairment and 7.3%, 11.9%, and 11.1% for people with mild neurocognitive disorder or dementia.

A multivariate model adjusted for age, gender, education, depression, HIV transmission category, CD4 count, viral load, exposure to antiretrovirals (including stavudine, didanosine, and indinavir), AIDS stage, hypertension, hypercholesterolemia, body mass index, and smoking status. In this model diabetes (compared with normal glycemia) was significantly associated with worse scores on six of the seven cognitive tests: Trail Making Test B of executive function (P = 0.01), Digit Symbol test of psychomotor speed (P = 0.0003), Purdue Pegboard test of manual dexterity (P = 0.0002), Rey complex figure test of visuospatial acuity (P < 0.001), Grober and Buschke episodic memory test (P = 0.0002), and Isaac Set test of semantic fluency (P = 0.02).

Compared with people who had normal glycemia, those with prediabetes had significantly worse scores on the Trail Making Test B (P = 0.04), the Digit Symbol test (P = 0.04), the Purdue Pegboard test (P = 0.02, and the Isaac Set test (P = 0.04).

A longitudinal analysis of annualized percent change in neurocognitive domains considered the same variables as the just-described cross-sectional analysis. Over time, scores declined significantly for people with diabetes (but not prediabetes or normal glycemia) in the Trail Making Test B, the Rey complex figure test, and the Grober and Buschke episodic memory test. Findings did not change after the researchers used inverse probability weighting to account for attrition from the cohort at the 2-year follow-up point.

The Aquitaine team concluded that HIV-positive people with diabetes perform worse on cognitive tests, especially those assessing executive function, attention, and psychomotor speed. They found no association between other cardiovascular risk factors and neurocognitive test scores.

The investigators noted that the mechanisms linking diabetes to neurocognitive function remain to be clarified, but they speculated that those mechanisms could involve the impact of HIV and its treatment on inflammation and microcirculation.

A case-control study of HIV-positive and negative US women without a history of diabetes found that higher levels of full-length soluble insulin receptor in plasma were associated with presence and severity of HIV-associated neurocognitive disorder (HAND) [2]. These researchers suggested that "insulin receptor dysfunction may have a role in the progression of HAND and could represent a biomarker for the presence and severity of HAND."

In a cross-sectional analysis of 130 US CHARTER cohort members, diabetes appeared to be associated with neurocognitive impairment only in older study participants [3].


1. Dufouil C, Richert L, Bruyand M, et al. Type 2 diabetes is associated with lower cognitive performances in a cohort of HIV-positive patients. ANRS CO3 Aquitaine Cohort, Bordeaux, France, 2007-2009. XIX International AIDS Conference. July 22-27, 2012. Abstract THAB0203.

2. Gerena Y, Skolasky RL, Velez JM, et al. Soluble and cell-associated insulin receptor dysfunction correlates with severity of HAND in HIV-infected women. PLoS One. 2012;7:e37358. http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0037358.

3. McCutchan JA, Marquie-Beck JA, Fitzsimons CA, et al. Role of obesity, metabolic variables, and diabetes in HIV-associated neurocognitive disorder. Neurology. 2012;78:485-492. http://www.ncbi.nlm.nih.gov/pubmed/22330412.