CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

infections (N = 332 controls) in Ho Chi Minh City between 2010 and 2015. Controls were matched to cases by age, sex, and CD4 count or WHO disease staging. We describe quantitative antigen dynamics over 90 days for 60 patients on treatment. Results: The median CD4 count was 18 cells/mm3 (IQR: 4-22) in cases and 52 cells/mm3 (IQR: 11-56) in control patients. Amongst cases, blood culture (automated Bactec system) was positive in 200 (70%) patients. The other 84 patients had positive culture from skin lesions or lymph nodes. Common opportunistic infections in control patients included cryptococcal meningitis (N=83), oral or esophageal candidiasis (N=47), tuberculosis (N=46), bacterial pneumonia (N=33), PCP (N=22) toxoplasmosis (N=22), and bacterial sepsis (N=21). Based on an optical density cut-off value generated by a receiver operating characteristic curve (ROC) of 0.208, the sensitivity, specificity, positive predicted value, and negative predicted value of the assay were 89.8%, 92.6%, 91.1%, and 91.5%, respectively. Quantitative antigen testing for 60 patients showed that 90% of patients had cleared antigenemia by 3 months. The rates of antigenemia clearance were similar for patients treated with amphotericin B or itraconazole as induction therapy (P=0.079, Student t-test). Conclusion: The Mp1p immunoassay provides an accurate test for differentiating talaromycosis from other HIV-associated opportunistic infections with sensitivity higher than blood culture, thus allows antifungal therapy to begin sooner, and has the potential to reduce talaromycosis mortality. 748 PREDICTING THE RISK OF DEATH IN HIV-ASSOCIATED TALAROMYCES MARNEFFEI INFECTION Thanh T. Nguyen 1 , Thu T. Nguyen 1 , Nguyen Le Nhu Tung 2 , Jeremy Day 1 , Cecilia Shikuma 3 , Chau V. Nguyen 2 , Guy Thwaites 1 , Thuy Le 1 1 Oxford Univ, Ho Chi Minh City, Vietnam, 2 Hosp for Trop Diseases, Ho Chi Minh City, Vietnam, 3 Univ of Hawaii, Honolulu, HI, USA Background: Disseminated Talaromyces marneffei infection (formerly termed penicilliosis) is the third most common microbiologically confirmed opportunistic infection in Southeast Asia with mortality of up to 30% despite antifungal therapy. There are no clinical algorithms to predict treatment outcomes. Methods: We performed logistic regression and developed a simple risk prediction model for HIV-associated talaromycosis using data from a retrospective cohort 2004-2009 (N=513) and a case-control study 2010-2011 (N=36) at the Hospital for Tropical Diseases in Ho Chi Minh City, Vietnam. Poor outcome was defined as death or worsening clinical status at hospital discharge. Covariables used in the model included age, sex, history of injection drug use, comorbidities, and clinical and laboratory characteristics. Based on the coefficients of predictors, a simple prognostic score was developed to estimate the risk of death. Single Conditional Mean Imputation was used for missing values of AST (22% missing) and creatinine (10%missing), and bootstrapping was used for internal validation of the final model. Results: 549 patients with microbiology-confirmed HIV-associated talaromycosis were included in the analaysis. Poor outcome was observed in 175/549 patients (31.9%). In the univariate logistic regression analysis, hepatomegaly and splenomegaly were protective factors. Shorter duration of illness, higher respiratory rates, dyspnea, AIDS-associated central nervous system (CNS) disease, platelet counts <50,000 cells/mL, AST >300 U/L, ALT >150 U/L, creatinine >110 µmol/L, and bacterial septicemia were predictors of poor outcome. In the multivariate logistic regression analysis, shorter days of illness (OR=1.22, 95% CI: 1.12-1.20, P=0.019), higher respiratory rates (OR=3.03, 95% CI: 2.08-4.41, P<0.001), CNS diseases (OR=18.91, 95% CI: 6.74–53.07, P<0.001), AST >300 U/L (OR=2.27, 95% CI: 1.22-4.21, P=0.009), and creatinine >110 µmol/L (OR=3.24, 95% CI: 1.82 – 5.74, P<0.001) were independent predictors of poor oucome. The prognostic scores ranged from -2 to +52, corresponding to a mortality risk of 0% to 100%. The internal validation showed acceptable discrimination (AUC=0.68) and calibration slope (1.00). The Brier score was 0.14. Conclusion: We developed a simple scoring system that can predict the risk of death in patients with HIV-associated talaromycosis based on routinely measured characteristics on admission. The scoring systemwill be externally validated using other cohorts in the region. 749 HOWWELL DO NEUROLOGIC SYMPTOMS IDENTIFY HIV-INFECTED INDIVIDUALS WITH NEUROSYPHILIS? Arielle P. Davis 1 , Joshua Stern 2 , Shelia Dunaway 1 , Lauren Tantalo 1 , Sharon Sahi 1 , Abigail Crooks 1 , Claire Stevens 1 , Sarah Holte 3 , Christina Marra 1 1 HarborviewMed Cntr, Seattle, WA, USA, 2 Univ of Washington, Seattle, WA, USA, 3 FRCHC, Seattle, USA Background: Current CDC guidelines recommend lumbar puncture in syphilis patients with signs or symptoms of neurologic disease. Methods: As part of a study of cerebrospinal fluid (CSF) abnormalities in syphilis, 410 HIV-infected individuals with untreated syphilis underwent lumbar puncture and a structured history that assessed the presence of new: headache, stiff neck, photophobia, vision loss, ocular inflammation, hearing loss, sensory loss, or gait incoordination. Severity was graded from 0 (absent) to 3 (severe). A symptomwas considered present if it was graded as ≥2 in severity. Neurosyphilis was defined as a reactive CSF-VDRL. Association between categorical variables was assessed by chi-square or fisher exact test and by logistic regression. P-values <0.05 were considered significant. Diagnostic specificity and sensitivity were calculated using standard formulae. Results: Participants were mostly white (73.4%) men (99.5%) with early syphilis (70.2%). Median RPR titer was 1:64 (IQR 1:16-1:256). Symptom frequency was: headache (18.5%), stiff neck (3.9%), photophobia (4.0%), vision loss (13.5%), ocular inflammation (4.4%), hearing loss (5.9%), sensory loss (0.7%) and gait incoordination (0.2%). CSF-VDRL was reactive in 69 (16.8%). Headache, stiff neck, photophobia and gait incoordination were not more common in those with a reactive CSF-VDRL. However, compared to those without each individual symptom, the odds of a reactive CSF-VDRL were significantly higher in those with vision loss (6.77 [95% CI 3.60-12.70], P<0.001) or hearing loss (3.28 [1.36-7.92], P=0.008); and bordered on significantly higher for those with ocular inflammation (2.58 [0.93-7.13], P=0.07) and sensory loss (10.03 [0.90-112.19], P=0.06). Taking into account serum RPR titer and antiretroviral use, the odds of a reactive CSF-VDRL remained significantly higher in those with vision or hearing loss. While the specificity of these 4 symptoms for neurosyphilis was high, the sensitivity was low (Table). Conclusion: In HIV-infected individuals with syphilis, new headache, stiff neck, photophobia and gait incoordination were not more common in those with neurosyphilis, while vision or hearing loss, ocular inflammation and sensory loss were. While the latter 4 symptoms had high specificity, they were very insensitive. Lack of neurologic symptoms in HIV-infected patients with syphilis should not reassure clinicians that their patients do not have neurosyphilis. 750 TREATMENT OF NEUROSYPHILIS: IV PENICILLIN G VS IM PROCAINE PENICILLIN/ORAL PROBENECID Shelia Dunaway 1 , Claire Maxwell 2 , Lauren Tantalo 1 , Sharon Sahi 1 , Arielle P. Davis 1 , Claire Stevens 1 , Abigail Crooks 1 , Christina Marra 1 1 HarborviewMed Cntr, Seattle, WA, USA, 2 Univ of Washington, Seattle, WA, USA Background: The CDC recommends IV aqueous penicillin G (PenG) for 10-14 days to treat neurosyphilis (NS). IM aqueous procaine penicillin (APPG) plus oral probenecid is considered an alternative treatment that might be considered if compliance can be assured. APPG is preferable to PenG in some instances due to lower cost and greater convenience. We compare these 2 therapies for neurosyphilis. Methods: Between April 2003 and May 2015, 150 individuals were enrolled in a study of cerebrospinal fluid (CSF) abnormalities in syphilis and were treated with PenG or APPG regimens for NS. They underwent follow-up CSF examinations and blood draws 3, 6, and 12 months after treatment. Relationships between categorical variables were determined by Chi square or Fisher exact test. Hazard ratios (HR) for normalization of CSF white blood cells (WBCs) (decline to <20/ul), CSF protein (decline to <50/mg/dl) and CSF-VDRL or serum RPR reactivity (4-fold decline or reversion to nonreactive) were determined using Cox regression. Results: Both groups were well matched; most were HIV-infected Caucasian men in their early 40s. More patients treated with APPG had early stage syphilis (71% vs 47%, p=0.01) and more were treated for uncomplicated syphilis within 90 days of study entry (43% vs 19%, p=0.01). There were no differences in pre-treatment serum RPR titer, CSF WBCs, CSF-VDRL reactivity, or proportion with symptomatic NS. More patients treated with PenG had elevated CSF protein (p=0.07). Normalization of all 4 measures did not differ in the 2 treatment groups, or in HIV-infected, or in those treated for uncomplicated syphilis before entry. CSF protein normalized more slowly in those with higher pre-treatment CSF concentration [HR=0.2, p <0.001]. CSF-VDRL normalized more slowly in late syphilis [HR=0.5, p <0.01] but normalized faster in symptomatic NS [HR=1.7, p <0.05]. Serum RPR normalized faster in those with higher pre-treatment titers [HR=2.3, p < 0.001] and in those with symptomatic NS [HR=1.5, p <0.05] and more slowly in late stage syphilis [HR=0.4, p <0.001]. Multivariate analysis, taking into account stage and previous treatment in addition to other significant univariate variables, revealed no relationship between normalization of any measure and treatment regimen.

Poster and Themed Discussion Abstracts

CROI 2017 325