Tackling Public Enemy #2: Fighting a deadly fungus Screening to Prevent Cryptococcal Deaths A common, deadly Weighing up strategies Screening to prevent Case studies and costly disease to prevent deaths deaths in South Africa Global burden of Fighting a deadly fungus HIV-associated cryptococcal meningitis 800,000 720,000 700,000 600,000 A common, deadly Weighing up strategies Screening to prevent Case studies 500,000 and costly disease to prevent deaths deaths in South Africa 400,000 300,000 200,000 120,000 Estimated yearly cases Estimated yearly 100,000 54,400 27,200 7,800 7,800 500 6,500 13,600 100 0 North Latin Caribbean Western North Africa Sub-Saharan Eastern South and East Asia Oceania America America and Central and Middle Africa Europe and Southeast Europe East Central Asia Asia Region ~1 million new cases per year and ~ 625,000 deaths per year Park BJ, et al. AIDS 2009;23:525-30. High burden of cryptococcosis in High, early mortality amongst South Africa adults accessing ART in sub-Saharan Africa Incidence of cryptococcosis (n=17,005*) vs. number of persons on antiretroviral treatment (ART)** by year, Gauteng Province, 2002-2010 Incidence 25 ART 300000 250000 antiretroviral treatment antiretroviral treatment 20 Numberon of persons 200000 15 150000 persons 10 100000 5 Incidence (cases per 100,000 per 100,000 Incidence (cases 50000 0 0 2002: 2003: 2004: 2005: 2006: 2007: 2008: 2009: 2010: n=1,194 n=1,511 n=1,539 n=2,000 n=2,253 n=2,109 n=2,141 n=2,141 n=2,117 Year *Complete surveillance audits were conducted throughout; **ASSA-2003 model Lawn S, et al. AIDS. 2008 1 3-pillared strategy Death from Cryptococcus in sub-Saharan Africa to reduce early mortality Estimated causes of death in sub-Saharan Africa, excluding HIV, 2009 Prevent, Early Strong screen HIV links & treat diagnosis to ART OIs Lawn S, et al. AIDS. 2008 Admitted to hospital Post-discharge survival in the during Mar-Nov 2002, and Jul-Sep pre-ART era 2003 High in-hospital mortality (n = 1,089) • Eight hospitals in Gauteng Induction treatment with amphotericin B and in-hospital case-fatality ratio for cases of incident • Follow-up post-discharge Disposition known Disposition lab-confirmed cryptococcosis (n=9,498*) diagnosed at GERMS-SA enhanced surveillance sites, – Pharmacy records (n = 1,037, 95%) unknown South Africa, 2005-2010 (n = 52, 5%) – Outpatient records 100 Amphotericin B Case-fatality ratio 40 – Interviews Death in hospital Discharged alive 90 (n = 316, 29%) (n = 721, 66%) 35 80 30 Case-fatality ratio (%) Case-fatality 70 25 60 Follow-up available Lost to follow-up 50 20 (n = 256, 36%) (n = 465, 64%) B B (%) 40 15 30 10 20 Cases treated with amphotericin treated Cases Last follow-up dead Last follow-up alive 5 10 (n = 154, 60%) (n = 102, 40%) 0 0 2005 2006 2007 2008 2009 2010 Died during readmission Died as outpatient Year to hospital (n = 86, 56%) (n = 68, 44%) Park BJ, et al. Int J STD AIDS. 2011. *Only includes cases with a CRF; missing treatment data for 226 cases and missing outcome data for 85 cases Post-discharge survival in the Admitted to hospital from Jan 2008 to Mar 2009 What does it cost annually to treat post-ART era (n = 255) patients in hospital? • One hospital in Gauteng • Follow-up post-discharge Met inclusion criteria Excluded – Pharmacy records (n = 217, 85%) (n = 38, 15%) Number of cases of – Outpatient records cryptococcal Cost of – Interviews Estimated annual cost Death in hospital Discharged alive meningitis per year X hospitalisation = (n = 60, 28%) (n = 157, 72%) 8,330 R 20,080 R 167,266,400 Follow-up available Lost to follow-up (n = 89, 57%) (n = 68, 43%) Last follow-up dead Last follow-up alive (n = 17, 19%) (n = 72, 81%) GERMS-SA Surveillance 2009; Haile et al. APHA Conference Atlanta, 2001 2 Fighting a deadly fungus Preventing deaths amongst patients with CD4 <100 A common, deadly Weighing up strategies Screening to prevent Case studies and costly disease to prevent deaths deaths in South Africa WHY IS SCREENING AN ATTRACTIVE OPTION? Cryptococcal Antigenaemia Prevalence in Published Studies 1 - Cryptococcal Antigen Country Year Setting Serum CrAg Notes Prevalence • Detectable in serum, plasma (& urine) before Zaire (Congo) 1989 Newly diagnosed HIV+ 12.2% (450) Includes symptomatic symptoms of meningitis develop Rwanda 1990 Laboratory serum tested 4.2% (213) Cross sectional • Average of 22 days prior to symptom onset1 South Africa 2011 ART enrollment hospital 3.0% (1033) No history of CM (Soweto) clinic CD4 <100 • Highly predictive of who is at risk for developing Prospective South Africa 2002-2005 Community clinic 7.0% (707) No history of CM cryptococcal disease (Cape Town) Retrospective CD4 <100 Uganda 2003-2004 ART clinic 5.8 % (377) CD4 < 100 • In Cape Town, 13/46 CrAg+ and 0/661 CrAg- patients Retrospective 2 developed meningitis Uganda 2004-2006 ART enrollment 8.2% (609) CD4 <200 Prospective • Prevalence of cryptococcal antigenemia ranges Uganda 2000 In and out-patients 10.7% (197) Stage III or IV Prospective from 3% to 21% Uganda 1995-1999 Two community clinics 5.6% (1372) • Highest amongst patients with CD4 <100 Cohort Cambodia 2004 ART enrollment 18.0% (327) Includes symptomatic Cross-sectional CD4 <50 Thailand 2008 Retrospective 9.2% (131) CD4 <100 1French et al. AIDS 2002; 2Jarvis et al. Clin Infect Dis 2009 2 - Missed opportunities for screening 2 - Missed opportunities for screening Admission to hospital with Antiretroviral treatment at time of diagnosis for cases of incident lab-confirmed Prior Prior initiation cryptococcal cryptococcosis (n=7,397) diagnosed at GERMS-SA enhanced surveillance sites, South diagnosis of of antiretroviral meningitis Death in Death post- Africa, 2005-2010 HIV infection treatment (n=1,468) hospital discharge On ART Not on ART 100% Timeline 90% 80% 70% 1,075/1,468 368/1,075 503/1,468 >50% of 60% (73%) (34%) (35%) persons who 50% survive admission 40% 30% 1 2 treatment antiretroviral Cases on 20% 10% 0% 2005 (n=954) 2006 (n=1,034) 2007 (n=1,247) 2008 (n=1,651) 2009 (n=1,647) 2010 (n=1,075) Year 3 3 – development of a new test for Lateral flow assay format Cryptococcus Antibodies conjugated to tag Test line Control line (gold, latex, fluorophore, etc.) (antibodies) (IgG antibodies) Analyte The new test (lateral flow Capillary flow assay) is: Sample Conjugate Nitrocellulose Wicking pad pad membrane pad Simple and quick Results available in 10 minutes Backing Available and effective Highly sensitive and accurate (>95%) Test line Control line Affordable (positive) (valid test) $2/test LFA can detect tiny amounts of LFA is very sensitive antigen in body fluids Serotype sensitivity (ng Crag/ml) Immunoassay Format A B C D IMMY LA 28 47 380 62 Meridian CALAS LA 19 37 940 54 Inverness LA 38 64 1600 50 Meridian Premier ELISA 28 23 >2000 770 mAbs F12D2 + 339 ELISA 0.6 0.8 5.0 0.6 IMMY LFA 1 1 9 8 Jarvis, et al. Clin Infect Dis. In press. Comparison of LFA vs. EIA 4 - An intervention exists… • Sets of samples from 62 patients with culture-proven cryptococcosis • Samples assayed by quantitative EIA and by LFA Serum Plasma Urine Serum Plasma Urine CrAg LFA (+) 61 61 61 Sensitivity 100% 100% 98% CrAg LFA (+/-) 1 1 0 95% CI 94-100% 94-100% 91-100% CrAg LFA (-) 0 0 1 Serum Plasma Urine 106 106 105 5 10 104 105 104 103 104 103 102 102 103 101 101 2 EIA - CragEIA - (ng/ml) 10 0 Correlation = 0.93 100 Correlation = 0.94 10 Correlation = 0.94 P < 0.001 P < 0.001 P < 0.001 101 10-1 10-1 101 102 103 104 105 106 100 101 102 103 104 105 106 10-1 100 101 102 103 104 105 LFA titer Meya D, et al. Clin Infect Dis. 2010 4 5 - Screening costs vs. cost of amphotericin B (based on prevalence of cryptococcal antigenemia) Preventing deaths amongst patients with CD4 <100 Meya D, et al. Clin Infect Dis. 2010 Fighting a deadly fungus Lab-driven screening strategy Reflex testing of remnant CD4 plasma A common, deadly Weighing up strategies Screening to prevent Case studies CD4 <100 and costly disease to prevent deaths deaths in South Africa National Screening Programme Cost Evaluation Cryptococcal Screening Programme Objectives Number of CD4 Cost of crypto LFA Estimated annual cost of specimens <100 test2 national screening programme 1 X = per year (R 21) R 13,356,000 (636,000) 1.Identify patients at risk (CD4 <100) Number of CM Cost of Estimated annual cost of 2.Test for cryptococcal antigen cases per year3 hospitalisation for current CM management X 4 = (8,330) CM R 167,266,400 3.Treat with oral fluconazole (R 20,080) 4.Prevent cryptococcal meningitis deaths Number of Cost of Estimated annual savings from preventable CM X hospitalization for = national screening programme cases per year5 CM R 96,384,000 (4800) (R 20,080) 1. NHLS Data Warehouse 2. Preliminary NHLS estimate 3. GERMS Surveillance 2009 4. Haile et al. APHA Conference Atlanta, 2001 5. Based on 3% CrAg+ positivity (Govender et al, unpublished) , 28% CM development among CrAg+ (Jarvis et al. Clin Infect Dis 2009), & 10% unpreventable deaths in pts presenting with overt CM (Meintjes, personal communication). 5 Proposed Cryptococcal Screening Algorithm for HIV+ Screening Fits into Routine HIV Care Patients Patient referred to HIV clinic for ART initiation Intake visit by nurse: CD4 & reflex CrAg (if CD4 <100) CRAG+ CRAG - Cryptococcal Screening Algorithm: No antifungal Antifungal treatment treatment Return visit to initiate ART at 1-2 weeks Decrease dose of † A lumbar puncture may considered in asymptomatic patients.