Cost-Effectiveness of a Lateral-Flow Urine Lipoarabinomannan Test for TB diagnosis in HIV-infected South African Adults Di Sun1; Susan Dorman2,3,4; Maunank Shah2,3,4; Yukari C. Manabe3,5; David W. Dowdy1,3,4 1 Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA; 2Johns Hopkins School of Medicine, Baltimore, MD, USA; 3Center for Tuberculosis Research, Johns Hopkins University, Baltimore, MD, USA; 4Tuberculosis Clinical Diagnostics Research Consortium (TB-CDRC); 5 Infectious Disease Institute, Kampala, Uganda

TB and HIV in South Africa WHO Global Tuberculosis Control 2011 People living with HIV/AIDS have increased risk of developing tuberculosis (TB) Rise in TB rates since 1990 and this is driven by the HIV epidemic Cite graph High early mortality among people with HIV-TB rapid diagnosis required to avoid death TB is the leading cause of non-injury death in South Africa TB diagnosis is especially difficult in HIV+ patients Sputum Smear Microscopy (SSM) sensitivity range as low as 20% Atypical clinical characteristics Novel modalities required to rapidly diagnose and treat these patients

Lipoarabinomannan (LAM) LAM is an immunogenic glycolipid found in the cell wall of Mycobacterium tuberculosis LAM antigen can be detected in urine Highest sensitivity in patients with high bacillary burden who have more detectable antigen in urine Immunosuppressed patients Disseminated TB Sensitivity is limited in individuals with CD4 >100 Can say that urine is easily accessible; Many HIV+ pts also have extrapulm TB http://www.birmingham.ac.uk/staff/profiles/biosciences/alderwick-luke.aspx

Lateral-Flow Urine LAM Assay Immunochromatographic point-of care test Requires minimal training No additional equipment or biosafety Cite the company Principle of the test Determine® TB LAM Ag is a lateral flow immunochromatographic rapid test for the qualitative detection of lipoarabinomannan (LAM) antigen in urine. The test is performed by applying 60 μL of urine and examining the reaction line 25 min thereafter. If LAM antigen is present in the sample, it binds to colloidal gold conjugated anti-LAM polyclonal antibodies and proceeds to bind to anti-LAM polyclonal antibodies on the capture line, forming a gray-purple line at the test window (marked patient). The results are stable up to 35 minutes. To assure assay validity, a procedural control line is incorporated in the assay device. Determine TB-LAM Alere, Waltham, MA, USA

Purpose Evaluate the cost-effectiveness of lateral-flow urine LAM assay in HIV-infected South African adults and the economic conditions under which it is most likely to be preferred

Methods Cohort study of hospitalized South African adults Cost-effectiveness analysis Decision analysis Primary outcome: Incremental Cost-Effectiveness Ratio ($/DALY averted) Sensitivity and uncertainty analyses performed on all parameters To determine the settings in which urine LAM may be cost-effective Date of evaluation for South African pts

Decision Analytic Model HIV+ (CD4<100) with clinical symptoms of TB Existing Diagnostics (CXR, SS, etc.) TB No TB Existing Diagnostics + LAM Assumed that at baseline all participants had an unspecified array of diagnostic tests etc. Comparing to added urine LAM testing to the same existing array of diagnostic tests Cost of these diagnostic tests the same in both arms

Decision Analytic Model: Base Case Scenario TB TB diagnosed by existing tests Treat Recover Die TB not diagnosed by existing tests Don’t Treat Assumptions: SS+ or LAM+ would definitely get treated A proportion of patients testing negative will also receive txt based on clinical judgment of treating physician This proportion remains constant regardless of availability of lateral flow urine LAM Undiagnosed, untreated TB has case fatality of 1.0 in this immunocompromised population

Existing Diagnostics negative for TB Decision Analytic Model: Base Case Scenario TB Existing Diagnostics (CXR, SS, etc.) indicate TB Treat Recover Die LAM indicate TB Existing Diagnostics negative for TB Don’t Treat Edit so that LAM doesn’t appear to occur after existing dx

Decision Analytic Model: LAM added TB Existing Diagnostics (CXR, SS, etc.) indicate TB Treat Recover Die LAM indicate TB Existing diagnostics negative for TB Don’t Treat Edit so that LAM doesn’t appear to occur after existing dx Existing diagnostics and LAM negative for TB

Parameter Values Name Value Range Reference TB Dynamics   TB Prevalence among individuals with HIV and CD4+ <100 cells/uL 0.38 0.12-0.5 Study data Probability of death in those with TB and given TB treatment 0.2 0.17-.23 WHO Characteristics of TB Diagnosis Probability of empiric treatment among smear-negative TB cases 0.53 0-0.75 Field et. al. Probability of empiric treatment among patients without TB 0.21 0-0.5 Martinson et. al. LAM Sensitivity 0.66 0.3-1 LAM Specificity 0.95 0.7-1 Sputum Smear Sensitivity 0.345 0.2-0.5 Lawn et. al. Sputum Smear Specificity 0.998 0.8483-1 Life Expectancy (yrs) HIV on ART (WHO Clinical Stage IV) 1.45 1-10 ASSA Unit Cost (2010 USD) TB Treatment $850 500-2000 LAM $3.50 2.98-30 Sputum Smear $1.58 1.34-1.82 Vassall et. al.

Existing Diagnostics + Cost-Effectiveness of Adding Lateral-flow LAM to Standard TB Diagnostics Cohort Size TB Cases TB Cases Treated False-Positives Treated DALYs DALYs averted Cost Incr. ICER $/DALY Existing Diagnostics 1000 380 262 130 495 $299,000 (ref) Existing Diagnostics + Urine LAM 342 155 437 58 $378,000 $79,000 $1370 Put in GDP per capita DALYS are so low b/c used conservative estimates.  still LAM was cost-effective  ICER < per capita gdp * Could also put the prob of ICER <GDP per capita Addition of urine lateral-flow LAM averts 58 DALYs at a cost of $1370 per DALY averted (95% uncertainty range: $710-3396). This is much less than the GDP per capita of South Africa ($7275).

Probability of cost-effectiveness May be able to take out this slide! Multivariable uncertainty analysis – varied all parameters to see whether urine LAM would fall under threshold Describe x & y axis * Per capita GDP is the WHO threshold for highly cost-effective interventions Falls to 86.3% at a threshold of $2000 and 33.6% at a threshold of $1000. The probability that urine LAM would cost less than the per-capita gross domestic product ($7275) was 98.3%.

Sensitivity Analysis $370 $1370 $2370 $3370 $370 $1370 $2370 $3370 Comment that the per capita gdp is way off the graph

Three-way Sensitivity Analysis Life Expectancy after TB Cure: 1.5 yrs Life Expectancy after TB Cure: 5 yrs Three way sensitivity analysis on prevalence, spec, and LE The black line is the set of parameters at which the ICER for urine LAM = per capita gdp for SA. Above line it’s more cost effective Mention ARTs! To be cost-effective at an assay specificity of 95% When life expectancy is 1.5 yrs, TB prevalence must be at least 5% When life expectancy is 5 yrs, TB prevalence must be at least 1%

Limitations Study outcomes: Sensitivity and Specificity No empirical evidence that addition of urine LAM improves survival Did not model transmission Transmission time unlikely to be reduced with addition of urine LAM due to advanced disease May not be generalizable to other populations Outpatient setting Other high-burden settings For high-burden settings- mention lower resources than SA

Conclusions Lateral-flow urine LAM is a feasible point-of-care test in hospitalized South African adults Urine LAM is a cost-effective diagnostic strategy ICER: $1370/DALY averted, South Africa GDP: $7275 Robust across wide range of sensitivity and uncertainty analysis Cost-effectiveness depends most strongly on LAM specificity, life expectancy, and TB prevalence Highly cost-effective with longer life expectancy Increasing life expectancy to 5 years, which is feasible with prompt antiretroviral agents, substantially improves cost-effectiveness In hospitalized, HIV-coinfected South African adults, addition of urine lateral-flow LAM testing to standard TB diagnostics detected an additional 80 cases of TB at an incremental cost of $79,000 per 1,000 individuals tested, or $1370 per DALY averted. Cost effectiveness- 50 years of life lived by one person = 1 year of life lived by 50; 1 yr of guaranteed life or 1/50 chance at 50 yrs

References WHO | Global tuberculosis control 2011. http://www.who.int/tb/publications/global_report/en/ (accessed 22 Feb2012). Salpeter SR. Fatal isoniazid-induced hepatitis. Its risk during chemoprophylaxis. West J Med 1993; 159:560–564. Field N, Murray J, Wong ML, Dowdeswell R, Dudumayo N, Rametsi L, et al. Missed opportunities in TB diagnosis: a TB process-based performance review tool to evaluate and improve clinical care. BMC Public Health 2011; 11:127. Martinson NA, Karstaedt A, Venter WF, Omar T, King P, Mbengo T, et al. Causes of death in hospitalized adults with a premortem diagnosis of tuberculosis: an autopsy study. AIDS 2007; 21:2043–2050. Lawn SD, Ayles H, Egwaga S, Williams B, Mukadi YD, Santos Filho ED, et al. Potential utility of empirical tuberculosis treatment for HIV-infected patients with advanced immunodeficiency in high TB-HIV burden settings [Unresolved issues]. The International Journal of Tuberculosis and Lung Disease 2011; 15:287–295. WHO | The global burden of disease: 2004 update. http://www.who.int/healthinfo/global_burden_disease/2004_report_update/en/index.html (accessed 2 Feb2012). Actuarial Society of South Africa. ASSA2008 AIDS and Demographic Model. http://aids.actuarialsociety.org.za/ASSA2008-Model-3480.htm (accessed 8 Feb2012). Vassall A, van Kampen S, Sohn H, Michael JS, John KR, den Boon S, et al. Rapid Diagnosis of Tuberculosis with the Xpert MTB/RIF Assay in High Burden Countries: A Cost-Effectiveness Analysis. PLoS Med 2011; 8:e1001120. Russell LB, Gold MR, Siegel JE, Daniels N, Weinstein MC. The role of cost-effectiveness analysis in health and medicine. Panel on Cost-Effectiveness in Health and Medicine. JAMA 1996; 276:1172–1177.

Acknowledgements Susan Dorman, MD David Dowdy, MD/PhD Yukari Manabe, MD Maunank Shah, MD Johns Hopkins Center for TB Research TB Clinical Diagnostics Research Consortium Doris Duke Clinical Research Foundation CDRC logo