Cepheid Symposium IAS 2017 Gwynn Stevens, PhD Director of Virology 24 July 2017

UNAIDS 90-90-90: Viral Load Essential Tool Delivering a better way to improve patient outcomes by enabling access diagnostic testing everywhere. Cepheid’s Mission is clear — at the end of the day, our purpose is to improve healthcare for all of us through fast, accurate diagnostic results that drive treatment decisions.

WHO Prequalification of In Vitro Diagnostics The aim of the WHO PQ DX is to promote and facilitate access to safe, appropriate and affordable diagnostics of good quality in an equitable manner WHO PQ undertakes a comprehensive assessment of the quality, safety and performance of IVDs Focus is placed on products for high burden diseases and their suitability for use in resource-limited settings. Cepheid’s values are an important part of who we are — it’s the backbone of our brand and the way in which we strive to do business.

Laboratory evaluation Prequalification decision WHO PQ Process Dossier review Site inspection Dossier incomplete  Laboratory evaluation Prequalification decision Dossier complete Dossier screening Priority product Yes No Pre-submission form

Performance Characteristics Analytical Performance Precision Linearity Limit of Detection Cross Contamination Clinical Performance Trueness of measurement – bias or mean difference Misclassification sensitivity and specificity Confidence intervals Discrepant results - >0.5log difference

Partners Analytical Study Clinical Study Oversight International Laboratory Branch, Division of Global HIV and TB CDC Atlanta Clinical Study NHLS HIV PCR LAB, Charlotte Maxeke Johannesburg Academic Hospital South Africa Oversight Prequalification of In Vitro Diagnostics Programme World Health Organization Geneva, Switzerland

Results: Analytical Study Reference Material

Results: Analytical Precision All %CV <3% Precision of measurement is acceptable

Results: Analytical Linearity Linearity verified in subtypes A, AG, B,C,D All slopes <5% from the ideal value of 1 R2 values all >99% - good correlation with the reference method Serial dilutions of stock panels, 7 members per dilution series 5 of each concentration Linear regression

Results: Limit of Detection (LOD) The claimed LOD was verified and estimated to be 22 copies/ml Lowest viral concentration detected with a positivity rate equal or higher than 95% WHO 3rd HIV-1 International Standard was used

Results: Cross Contamination No cross-contamination was detected 20 Subtype C alternating with 20 HIV neg

Results: Clinical Study Trueness of Measurement N=439 EDTA Plasma samples Correlation is within range (R2=0.941, P<0.001) Bias or mean difference 0.043 log10 copies/mL (95% CI: -0.986; 1.073) Trueness of measurement using Bland Altman Upper and lower limits of agreement, 2SDs, CI

Results: Clinical Study Misclassification Overall invalid/error rate 2.94% Sensitivity (95% CI): 225 /239= 94.14 % (90.37-96.76) Specificity (95% CI):197/200 = 98.50 % (95.68-99.69) False treatment failures: 3/ 200 = 1.50% False treatment successful: 14/239 = 5.86%

Conclusion Xpert HIV-1 VL Assay successfully meets all WHO prequalification requirements

Questions?