Elodie L. Plan, Kyle T. Baron, Marc R. Gastonguay, Jonathan L. French, William R. Gillespie, and Matthew M. Riggs
Metrum Research Group & Metrum Institute, Tariffville, CT, USA
Objectives: A physiologically-based multiscale systems (PBMS) model has been used to describe therapeutic effects in postmenopausal (PM) osteoporosis and other related diseases [1, 2]. This model has been expanded to describe bone-related effects associated with estrogen loss during menopause transition [3]. An extension to include fracture risk would broaden its utility in predicting outcomes affected by disease progression and drug intervention. Fracture risk has previously been associated with bone mineral density (BMD) [4]; however, studies rarely exceed 6 y in observational period length or consider menopause as the onset of the disease.
Our aim was to develop a model simultaneously characterizing BMD and fracture risk based on time since final menstrual period (FMP), in order to extend the PBMS model.
Methods: Publicly available NHANES datasets (demographics, dual energy X-ray absorptiometry, and osteoporosis questionnaire) [5] were accessed to combine data from 1439 PM women, each contributing 1 femoral neck BMD measurement at the time of examination. All non-trauma fractures reported since the age of 50 were included and represented 137 events over a maximum observation range of 35 y. A piecewise linear regression model based on a recent study [6] was applied to predict BMD time-course relative to FMP for each woman. A repeated time-to-event (RTTE) model [7] including time-varying BMD as a predictor was fit to the fractures. The BMD and RTTE analyses were performed simultaneously in WinBUGS 1.4.3 with a Component Pascal ODE specification and a LSODA solver [8].
Results: The estimated population BMD baseline was 0.8 g/cm2, which is consistent with other studies. The profile was satisfactorily described; obtained transmenopausal, PM, and final slopes were -1.8, -1.2, and -0.3%/y, respectively. Covariate effects (BMI, ethnicity, and FMP age) were of similar magnitude than previously reported. The initial posterior median of the fracture hazard was quantified as 0.007 (95% credible interval 0.004-0.011); the associated random effect significantly decreased once the BMD effect was introduced. The RTTE component was integrated into the PBMS model, varying with BMD expressed here as a turnover model linked to osteoblast and osteoclast functions, allowing for age and FMP-related estrogen declines to be associated with changes in fracture rates.
Conclusions: The addition of fracture probability to the PBMS model enables the characterization of a primary endpoint used in osteoporosis clinical trials.
References:
[1] Peterson MC and Riggs MM. Bone, 2010 Jan;46:49–63.
[2] Riggs MM, Peterson MC, and Gastonguay MR. J Clin Pharmacol, 2012 Jan;52:45S–53S.
[3] Riggs MM, Gillespie WR, Gastonguay MR, and Peterson MC. National Institute of General Medical Sciences Quantitative Systems Pharmacology Workshop II, Bethesda, MD, 2010 Sep.
[4] Garnett C and Holford NHG. 5th International Symposium on Measurement and Kinetics of In Vivo Drug Effects. Noordwijkerhout, the Netherlands, 2006 Apr 26-29.
[5] Centers for Disease Control and Prevention (CDC). National Center for Health Statistics (NCHS). National Health and Nutrition Examination Survey Data. Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, [2005-2006,2007-2008] [http://www.cdc.gov/nchs/nhanes/nhanes_questionnaires.htm].
[6] Greendale GA, Sowers M, Han W, Huang MH, Finkelstein JS, Crandall CJ, Lee JS, Karlamangla AS. J Bone Miner Res, 2012 Jan;1:111-118.
[7] Cox EH, Veyrat-Follet C, Beal SL, Fuseau E, Kenkare S, Sheiner LB. J Pharmacokinet Biopharm, 1999 Dec;27(6):625-44.
[8] BUGSModelLibrary, PKPD Model Library for WinBUGS, Metrum Institute [http://code.google.com/p/bugsmodellibrary/]
Reference: PAGE 21 () Abstr 2592 [www.page-meeting.org/?abstract=2592]
Poster: Other Modelling Applications