M. Prague (123), C. Pasin (4), R. Thiébaut (123) & ANRS CO3 Study group
(1) Inria Bordeaux, SISTM Team, Bordeaux, France (2) Inserm U1219 Bordeaux Public Health, SISTM Team, Bordeaux France (3) Vaccine Research Institute, Creteil, France (4) Department of Pathology and Cell Biology, Columbia University Medical Center, New York, New York, United States of America
Objectives: The challenge associated with lifelong combination antiretroviral treatments (cART) taken by HIV infected patients have motivated the development of strategies for therapeutic relief. The cART simplification reduces toxicity and drug costs, and improves patients’ quality of life. Short-cycles treatment interruptions (SCT), in which patients take their treatment only few consecutive days in the week, have been shown as promising and are still under investigation. The PENTA Trial group investigated 5/7 designs (5 days on, 2 days off cART) in adolescents in the Breather trial. They showed a sustainable non-inferiority of virologic suppression compared to continuous cART [1]. The ANRS 170 QUATUOR is an ongoing trial investigating 4/7 designs and conclusions are yet to be drawn We propose a pipeline for computer-based simulations of trials which aim at quantifying and predicting in silico the effect of ongoing and future SCT in order to accelerate and personalize their development.
Methods: HIV viral load and CD4+ T cells count trajectories under treatment conditions were modelled using dynamical mechanistic models based on Ordinary Differential Equations [2,3]. Data from 2550 patients of the ANRS CO3 Aquitaine HIV cohort were used to estimate model parameters. Because longitudinal observations were sparse and prone to error measurement, we adopted a population mixed effects approach and supplemented it with information from in vitro assays. Of particular interest, we used Instantaneous Inhibitory potentials (IIP) of antiretroviral and cART to increase information [4]. Estimates of the in vivo cART effect were obtained from a populational statistical optimization approach using both NIMROD [5] and Monolix [6]. We generate pseudo subjects sampled with statistically controlled heterogeneity, for whom pseudo parameters are sampled from the estimated parameters distribution. Then, by predictions, we simulate in silico data of realistic SCT trials. Computer-based success of the SCT is evaluated regarding the distribution of the probability of detectable viral load (at level 1 and 50 copies/mL) and the distribution of the basic reproduction number R0 at 48 weeks in the in silico population. The R0 can be thought of as the number of d CD4+ T infected cells one infected cells generates on average over the course of its infectious period, in an otherwise uninfected CD4+ T cells population. If R0 is below 1 the infection will die out in the long run. Therefore, R0 provide additional mechanistic information on the computer-based probability of success of the SCT.
Results: We estimate that most of the investigated Efavirenz-based (EFV) cART will be potent enough to guarantee the success of 5/7 designs (5 days on, 2 days off cART). We can derive in silico the results of the Breather trial [7], which is a 5/7 design for EFV-based cART. Simulations predict 1% [0%; 11.9%] of virologic failure and show a mean R0 of 0.82 [0.63; 0.99]. The in silico and in vivo results were consistent as the 95% confidence interval included the true outcome (8.1% of 50 copies/mL virologic failure at 48 weeks). The investigation of 4/7 designs show that with the most potent cART (such as dolutegravir- or darunavir- based cART), viral suppression is likely to be sustained, simulations predict between 0% and 2% of virologic failure.
Conclusions: The computer-based approach correctly predicts the outcome of existing SCT trials. Moreover, it has the advantage to characterize mechanistic indicators, such as the R0, which provides alternative outcome for the SCT success. Thus, our pipeline for in silico trials is a promising tool for accelerating the development of novel strategies based on existing cART. Finally, as a direct extension of the methodology, it is also a promising tool to investigate individual- and treatment- specific SCT.
References:
[1] Turkova, A., Moore, C. L., Butler, K., Compagnucci, A., Saïdi, Y., Musiime, V., … & Scott, K. (2018). Weekends-off efavirenz-based antiretroviral therapy in HIV-infected children, adolescents and young adults (BREATHER): Extended follow-up results of a randomised, open-label, non-inferiority trial. PloS one, 13(4), e0196239.
[2] Perelson, A. S., & Deeks, S. G. (2011). Drug effectiveness explained: the mathematics of antiviral agents for HIV. Science translational medicine, 3(91), 91ps30-91ps30.
[3] Prague, M., Commenges, D., Gran, J. M., Ledergerber, B., Young, J., Furrer, H., & Thiébaut, R. (2017). Dynamic models for estimating the effect of HAART on CD4 in observational studies: Application to the Aquitaine Cohort and the Swiss HIV Cohort Study. Biometrics, 73(1), 294-304.
[4] Jilek, B. L., Zarr, M., Sampah, M. E., Rabi, S. A., Bullen, C. K., Lai, J., … & Siliciano, R. F. (2012). A quantitative basis for antiretroviral therapy for HIV-1 infection. Nature medicine, 18(3), 446.
[5] Prague, M., Commenges, D., Guedj, J., Drylewicz, J., & Thiébaut, R. (2013). NIMROD: A program for inference via a normal approximation of the posterior in models with random effects based on ordinary differential equations. Computer methods and programs in biomedicine, 111(2), 447-458.
[6] Monolix version 2018R2. Antony, France: Lixoft SAS, 2018.
http://lixoft.com/products/monolix/
[7] The Breather Trial Group (2016). Weekends-off efavirenz-based antiretroviral therapy in HIV-infected children, adolescents, and young adults (BREATHER): a randomised, open-label, non-inferiority, phase 2/3 trial. The lancet HIV, 3(9), e421-e430.
Reference: PAGE 28 (2019) Abstr 8901 [www.page-meeting.org/?abstract=8901]
Poster: Drug/Disease Modelling - Infection