Periprosthetic bone loss after cementless THA is a well-cited phenomenon and may contribute to the longevity of implant survivorship. Periprosthetic bone mass decrease after joint replacement is a continuous process and has been reported to last for over 10 years.[17] At the proximal femur, loss of bone stock has been reported to account for 16–30%.[18] This phenomenon increases the risk of early implant migration, aseptic loosening, periprosthetic fracture, and hence decreased implant survivorship. [4, 19]
BMD value is an important factor to the fixation and stability of the implanted prosthesis.[20] Early stem migration and rotation has been detected in osteoporotic bone compared to bones of normal BMD values in patients undergoing THA.[21] Patients with low preoperatively BMD values are at risk of major complications, including stem migration or even subsidence.[22] Currently, many studies support the role of bisphosphonate in the preservation of periprosthetic bone after THA.[23, 24] ZA, a third-generation bisphosphonate, has shown to be more potent than its second-generation predecessor in multiple in-vivo studies.[25–27] In our previous study, a short term use of ZA effectively increased periprosthetic BMD values.[11] In this five year extension study, ZA group retained higher BMD values across all seven Gruen zones. Although BMD levels at the fifth year were not statistically different between ZA group and control group, fifth year BMD change ratios in Gruen zones 2, 4, and 6 were significantly higher in ZA group as compared to the control group. Parallel to the above findings, the levels of BTMs (ALP, OC, and P1NP) were shown to be persistently suppressed in ZA group after ZA discontinuation for 4 years. In our previous study that included 54 patients, the baseline OC and P1NP levels were not different between groups. However, the 25-patient ZA group within the 49-patient extension study had significantly higher baseline levels of OC and P1NP (p = 0.044 and p = 0.030, respectively) as compared to the control group, suggesting a higher pre-trial bone turnover status. Nevertheless, ZA given one day after THA could potently inhibit P1NP and ALP levels in ZA group as early as 6 weeks, and booster dose of ZA at one year after THA further inhibited P1NP, ALP and OC levels at 2 years or 5 years. High bone turnover status with increased serum OC and P1NP is associated with low BMD in osteoporotic population. [28, 29] OC and P1NP are also useful BTMs for monitoring antiresorptive therapy and fracture risk assessment. [30–32] We found 2 doses of ZA given within one year after THA had lasting effects on the inhibition of P1NP/OC at 2 years and ALP/OC at 5 years. To the best of our knowledge, the interesting findings of the lasting effect of ZA on periprosthetic BMD and bone turnover after drug holiday for 4 years has not been reported. It may be related to ZA's high affinity to bone and high potency pharmacological nature such as seen in the extension study of HORIZON pivotal fracture trial.[33] Reports of other bisphosphonates (risedronate, alendronate, or pamidronate) have shown good periprosthetic bone preservation in early follow-up but not in mid-term [19, 23, 34, 35] as the increased periprosthetic BMD would eventually decrease after discontinuation of these aforementioned medications. Some authors have thus suggested that lifelong administration of bisphosphonate may be warranted, especially for patients with poor preoperative bone stock.[36, 37] Our results are in accordance with other studies on ZA drug holiday, as it demonstrates a prolonged effect after discontinuation.[38] Findings in this study show that a 4-year drug holiday of ZA can preserve periprosthetic BMD, inhibits BTMs, and exhibit no adverse reactions such as the rebounding phenomenon seen in RANKL antagonist withdrawal [39, 40], ONJ, or atypical bony fractures. However, no difference was found in regards to the functional outcomes and implant survival. All implants are well fixed without loosening, migration, or pedestal formation. No patients received revision surgery for periprosthetic joint infection or periprosthetic fracture. Nevertheless, we find the protective potential of a short-term administration of ZA lasting up to 5 years encouraging.
This study has limitations. Firstly, the sample size was small, with the initial study including 60 THA's randomized to ZA and placebo. At the fifth year, 49 recruits completed the clinical follow-up while only 19 cases had completed BMD study. Secondly, randomization of open label ZA or placebo was not blinded by a dummy for control subjects. Thirdly, while good patient compliance was achieved with no other anti-osteoporotic medication administered during the study period, calcium and vitamin D supplements were not strictly controlled. Finally, this study reports the cross-sectional results at the fifth year in a cohort receiving 2 doses of ZA and subsequent drug holiday for 4 years. These results may not be translated onto other dosing regimen or projected for long-term outcomes. However, there are merits in this extension study in that data of the prospective randomized clinical trial is unbiased and unique, and all surgeries were performed by an experienced surgeon. The study was stringently performed and audited by the institutional research board.
In conclusion, this study demonstrates the lasting effect of a two-dose ZA given within one year after THA on bone metabolism and periprosthetic BMD at 5 years. The short-term dosing of ZA and drug holiday for 4 years has no adverse events and results in significant inhibition of periprosthetic bone loss and bone turnover markers.