Validation of the EACS guidelines 2011 algorithm for detecting HAND in the CHARTER cohort.

Type: Poster
Title: Validation of the EACS guidelines 2011 algorithm for detecting HAND in the CHARTER cohort.
Authors: Perez-Valero I, Heaton R, Letendre S, McCutchan J, Morgello S, Gelmen B, Collier A, Grant I, and the CHARTER Group
Date: 03-05-2012
Abstract:Background: The European AIDS Clinical Society (EACS) published in October 2011 guidelines for the detection and management of HIV-associated neurocognitive disorders (HAND). EACS guidelines use self-reports of cognitive difficulties and were exclusively focused on screening for patients at-risk for symptomatic HAND (sHAND). The current analyses were conducted to evaluate the accuracy of the EACS screen in detecting sHAND and to compare this with the accuracy of a performance-based screen, the HIV dementia scale (HDS). Methods: This cross-sectional analysis included all participants of the CHARTER cohort who did not have confounding comorbidities (n = 1341). Symptomatic HAND included all MND and HAD diagnoses. Non-sHAND included subjects with asymptomatic HAND (ANI) and subjects without HAND. Gold standard diagnostic classifications were based upon CHARTER’s comprehensive neurobehavioral assessments. EACS screen was defined positive when at least 1 of 3 questions extracted from the Patient’s Assessment of Own Functioning Inventory were positive, plus report of increased dependence in at least one everyday activity. An HDS result <14 was considered positive. Screening test comparisons were performed using likelihood ratios (LR) to avoid the impact of prevalence. A secondary analysis evaluating the accuracy of the 2 screens for detecting all HAND diagnoses (both symptomatic and asymptomatic) was also performed. Results: Six hundred and twelve participants with HAND (ANI 457, MND 122, and HAD 33) and 729 without HAND were included. The EACS screen was positive in 153 participants, 89 with sHAND and 64 with non-sHAND. HDS was positive in 504 participants, 80 with sHAND and 424 with non-sHAND. Classification accuracies of both screens are reported in the table. EACS screen was superior to HDS in identifying sHAND. However, the HDS was superior in detecting the combined asymptomatic and symptomatic forms of HAND (see table and figure). Conclusions: Diagnostic screens aim to identify cases at risk for HAND, resulting in referral for more comprehensive assessments; as such, sensitivity to the screened condition is critical. Neither EACS nor HDS screens show adequate sensitivity for this purpose. EACS sensitivity is especially poor if the full range of HAND is considered.