CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

714 DIGITAL CHEST RADIOGRAPHY OPTIMISES TB SCREENING OF SOUTH AFRICAN CLINIC ATTENDEES Nishila Moodley 1 , Amashnee Saimen 1 , Noor Zakhura 2 , David Motau 2 , Candice Chetty-Makan 1 , Kavindhran Velen 1 , Gavin Churchyard 1 , Salome Charalambous 1 1 The Aurum Institute, Johannesburg, South Africa, 2 South African National Department of Health, Pretoria, South Africa Background: In 2016, it is estimated that approximately 165 000 Tuberculosis (TB) cases were missed in South Africa. Optimising TB screening is imperative in meeting national and global TB targets, including identifying 90% of all TB cases. Digital CXR (d-CXR) provides a quick, reproducible technique incurring lowmarginal costs, reduced radiation exposure and improved portability for TB screening. We assessed whether d-CXR screening with Computer-Aided Detection for Tuberculosis (CAD4TB) would improve TB yield when combined with the World Health Organisation (WHO) TB 4-question symptom screening tool. Methods: A systematic sample of adult patients attending three public health clinics for any reason (excluding ante-natal care) in the Free State Province, South Africa, were screened for TB between November 2017 and June 2018 using d-CXR and the WHO TB symptom tool. Patients <18 years, pregnant, currently receiving or received anti-tuberculosis treatment within the past two years were ineligible for participation. Two spot sputumwere collected for Xpert MTB/RIF Ultra assay (Xpert Ultra) and MGIT culture from attendees with ≥1 TB symptom and/or a CAD4TB score of ≥60. All participants were offered HIV testing. TB yield was compared between screening strategies and the number needed to test (NNT) determined. Results: We approached 4352 clinic attendees, 3.8% refused participation, 26.3%were ineligible and 3,041 participants were screened (2,005 female [65.9%], mean age 45 years (SD 15.2), HIV prevalence 36.3% [1,030/2,837]). The proportion of attendees screened by d-CXR, symptoms and d-CXR/symptoms requiring TB investigations was 19% (573/3041), 36% (1109/3014) and 45% (1356/3041) respectively. The yield of TB (Xpert Ultra/culture positive) for d-CXR, symptom and d-CXR/symptom screen was 2.2%, 2.3% and 2.8% and the NNT was 8.7, 15.4 and 16.1 respectively. Conclusion: A high proportion of clinic attendees had symptoms suggestive of TB. The addition of d-CXR to symptom screening improved TB yield with a modest increase in the number requiring TB investigations. D-CXR alone compared to symptom screening alone had a similar yield of TB and almost halved the number requiring TB investigations. D-CXR screening alone is potentially a cost effective TB screening strategy.

713 PNEUMOCOCCAL VACCINATION IN HIV+ ADULT PATIENTS ON SUPPRESSIVE ART, 2010-2017

María J. Vivancos-Gallego , Alfonso Muriel, Ana Moreno, Maria Jesus Perez- Elias, Jose L. Casado, Carmen Quereda, Matilde Sanchez-Conde, Santos Del Campo, Fernando Dronda, Sergio Serrano-Villar, Ana Sanchez, Jose Valencia, Santiago Moreno, for the Instituto Ramon y Cajal de Investigacion Sanitaria Hospital Ramón y Cajal, Madrid, Spain Background: There is little information on the efficacy of the pneumococcal vaccines (PV), especially of the pneumococcal conjugate vaccines (PCV), in successfully treated patients in the modern ART era. Methods: Case-control study in a tertiary, University Hospital in Madrid. Cases were HIV-patients admitted to the hospital (2010-2017) with a microbiologically confirmed infection due to S. pneumoniae (from a normally sterile site and/or a positive urinary antigen). Controls (HIV-infected patients without IPD) were selected by random sampling matched with cases by gender and year of HIV diagnosis. The selection was blind for the study factor (vaccination). Sample size was estimated (61 cases and 183 controls). We performed comparisons to vaccine exposure and outcome associations using time-dependent covariates in a Cox proportional-hazards regression model. Results: The population of study included 256 subjects, 64 cases and 192 controls. Male 77%, median age 29, previous AIDS 43%, median Charlson Comorbidity Index 6. 115 (45%) patients had been vaccinated. Median CD4-cell count at the time of administration of the PV was 518 (318-733) cells/mL, 79% with HIV RNA<50. In a multivariate logistic regression analysis, risk factors associated with IPD were the Charlson Comorbidity Index (HR 1.23 95%CI 1.14-1.33 P=0.0001) and previous diagnosis of AIDS (HR 2.82 95%CI 1.26-6.33 P=0.012), while plasma HIV RNA <50 copies/mL (HR 0.44 95%CI 0.22-0.86 P=0.016) was protective. PV was not associated with IPD after adjusting in the multivariate model with time protection as a dependent covariate (HR 0.64, 95% CI 0.33-1.25 P=0.191). We also investigated the influence of different PV schedules. In univariate analysis, compared to no vaccine, no significant protection was found in patients who received only PPSV-23 or only a conjugate vaccine (PCV-7 or PCV-13), while two vaccines given in series (PCV13- PPSV-23 or PPSV23-PCV13) showed protection (HR 0.3 95% CI 0.11-0.77 p=0.012). However, in an adjusted model we found no evidence of protection by double PV schedules (HR 0.44 95%CI 0.17-1.14 p=0.09). Conclusion: In this case-control study, different schedules of pneumococcal vaccination did not show protection against IPD. As only plasma HIV RNA <50 copies/mL was found to be a protective factor, early ART initiation could ensure the protection in most patients. As with HIV-uninfected persons, the pneumococcal vaccination should then be individualized in HIV-infected patients based on traditional risk factors for IPD.

Poster Abstracts

CROI 2019 274