Skip to main content
. 2011 Dec 6;343:d7222. doi: 10.1136/bmj.d7222

Table 2 .

Parametric Weibull model assessment of the effect of bisphosphonate use on prosthesis survival, up to 15 years after surgery

Study group Unadjusted models Propensity adjusted models
Hazard ratio (95% CI) for implant failure Time ratio (95% CI) for implant survival P value* Hazard ratio (95% CI) for implant failure Time ratio (95% CI) for implant survival P value*
Total population (n=41 995) 0.41 (0.22 to 0.76) 2.54 (1.34 to 4.83) 0.004 0.54 (0.29 to 0.99) 1.96 (1.01 to 3.82) 0.047
Population diagnosed with osteoarthritis (n=34 336) 0.30 (0.13 to 0.70) 3.47 (1.44 to 8.37) 0.005 0.40 (0.17 to 0.93) 2.69 (1.07 to 6.73) 0.034
Population undergoing knee arthroplasty (n=18 726) 0.33 (0.12 to 0.88) 2.79 (1.12 to 6.91) 0.027 0.40 (0.15 to 1.07) 2.37 (0.94 to 6.01) 0.068
Population undergoing hip arthroplasty (n=23 269) 0.47 (0.24 to 0.94) 2.36 (1.07 to 5.21) 0.033 0.64 (0.32 to 1.28) 1.71 (0.74 to 3.94) 0.20

P values apply to both time ratios and hazard ratios.