This propensity-matched cohort study represents the largest study examining the association between reoperation or periprosthetic fracture and osteoporosis after primary TAA. Contrary to our hypothesis, the present study findings suggest that osteoporosis is not a significant risk factor for reoperation or postoperative periprosthetic fracture through 3 years following primary TAA.

In this study, the incidence of reoperation 3 years following primary TAA was 5.9% in the osteoporosis cohort and 5.6% in the non-osteoporosis cohort. When combined, the overall reoperation rate across both cohorts was 5.2%, which aligns with the expected values from the literature on patients over the age of 50[10, 42, 43]. On this topic, Demetracopoulos et al. conducted a comparative study involving 395 TAA recipients sub grouped by age (i.e., < 55, 55–70, > 70 years). After a mean follow-up of 3.5 years, the incidence of revision among those aged 55–70 years was 5.4% and not significantly different compared to the other age groups.[42] Although older individuals in this study may have had a greater prevalence of osteoporosis, the authors did not directly compare bone quality across the three groups, thereby limiting any specific insights into the association between osteoporosis and TAA outcomes[21].

Only one previous small chart review examined the relationship between BMD and TAA outcomes, and found no association between BMD and TAA revision, yet found a positive association between lower BMD and periprosthetic fracture[28]. In this study, including 30 TAA recipients with Housenfield unit (HU) measurements derived from computed tomography of the tibia, Cody et al. found that lower tibial HU values, which correspond with lower BMD, demonstrated a positive significant association with intraoperative and postoperative periprosthetic fracture.[28] There was no association between HU and TAA revision over a median of 2.4 years’ follow-up. However, Cody et al. only controlled for age, sex, and weight, and therefore the study findings are limited by potential unmeasured confounding related to comorbidities associated with complications of TAA, such as diabetes, smoking, and inflammatory arthritis.[18,19,20,21] In addition, it was unclear how many patients in this study had osteoporosis. Differences between the markers used for low BMD in the study by Cody et al. and our present study may also explain the discrepancy compared to the present study regarding the likelihood of periprosthetic fracture. While HU is a continuous measure, a diagnosis of osteoporosis is a binary distinction based on a defined threshold of BMD based on the results of DXA, according to the World Health Organization (T-score of ≤ 2.5). While HU has been used widely for assessing bone quality, DXA remains the gold standard for clinically diagnosing osteoporosis.[44] Accordingly, our population appears more selective, potentially representing osteoporotic patients with a comparatively lower BMD than those with periprosthetic fracture in the Cody et al. study, who had a mean HU value of 204 (SD = 113). As HU values less than 100 are typically considered to represent osteoporosis, while those over 160 are considered normal,[45] the mean HU values among those with fracture in the Cody et al. study fell within a range greater than would be expected for osteoporosis.

Osteoporosis disproportionately affects women, particularly postmenopausal women, due to the rapid decline in estrogen levels, which plays a crucial role in maintaining BMD. In contrast, men experience a more gradual loss of BMD with aging, often driven by declining testosterone levels[46]. Despite the lower prevalence of osteoporosis in men, fractures in this population tend to result in greater morbidity and mortality compared to women[47, 48]. This disparity underscores the need for targeted screening and intervention strategies tailored to each gender. While this study did not stratify outcomes by gender, future research should investigate whether sex differences influence surgical outcomes, such as reoperation or periprosthetic fracture rates, after TAA.

In general, osteoporosis is highly prevalent in older adults undergoing arthroplasty[49] and contributes to the increasing incidence of periprosthetic fracture and aseptic loosening after primary total knee or hip arthroplasty[22, 24, 50,51,52]. In a recent investigation of a national administrative claims database of 418,054 patients by Harris et al., of which 10% had osteoporosis, the 5-year likelihood of all-cause revision after TKA after controlling for age, sex, and the Charlson comorbidity index was slightly higher for patients who had osteoporosis (hazard ratio = 1.1, 95% CI: 1.0, 1.2). Further, in that study, osteoporotic patients had an approximately twofold increased risk of 5-year revision for periprosthetic fracture after TKA after controlling for the aforementioned patient characteristics.[53] While it is unknown why these results for TKA differ than those from TAA in our study, differences may be attributed to implant type or joint mechanics, among other factors. Despite this, preoperative screening with DXA scan remains underutilized and thus the impact on osteoporosis remains underdeveloped[54]. However, there is growing evidence that bisphosphonate treatment pre- and postoperatively may have protective effects, as these medications have been shown to preserve periprosthetic BMD for more than 5 years after THA[55, 56]. Furthermore, in large propensity score analysis study of patients undergoing TKA, bisphosphonate use postoperatively was associated with a 50% reduction in fracture risk[57]. However, it is currently unknown if bisphosphonate or other osteoporosis medication use, such as denosumab, could contribute to a reduction in fracture risk in patients undergoing TAA.

In contrast to the hip and knee arthroplasty literature, our findings suggest that patients with osteoporosis are not at increased risk of requiring revision TAA or experiencing a periprosthetic fracture. The potential impact of osteoporosis on TAA might be counterbalanced by the protective effects of surgical techniques and implant technologies. Accordingly, intraoperative assessment of bone quality remains essential for surgical decision-making during TAA, regardless of the results of this study. Prophylactic fixation of the medial malleolus is advisable in patients with poor bone quality given the risk of medial malleolar stress fracture[58]. Further, increased bony support has been shown to reduce implant and bone micromotion and improve implant stability and decrease mechanical failures. Thus, when securing the implant components in the presence of deficient bone stock and instability around the ankle joint, the adoption of a modular long-stem tibial component and a talar implant reinforced by two talar pegs can enhance stability and reduce the risk of loosening[2]. In addition, patients with osteoporosis may have been more commonly treated with long-stem tibial components relative to patients without osteoporosis. Regardless, while robust tibial fixation may serve as a risk mitigation factor and may be the reason that osteoporosis was not associated with reoperation or periprosthetic fracture in this study, perioperative assessment and medical optimization of low bone density should be further investigated as a possible strategy to improve patient outcomes.

Our study exhibits both strengths and limitations that warrant consideration. Strengths include adherence to a registered protocol, inclusion of a multidisciplinary author team, a relatively large sample size, and robust propensity matching strategy. We were unable to stratify patients according to bone mineral density, or the bone density at the operative site. While these variables may confound the results, DXA scans are not routinely performed prior to TAA, and typically only include measurements of the hips and lumbar spine rather than the ankle. Furthermore, implant type, concomitant procedures, and surgeons’ level of experience may have represented unmeasured confounders. It was not feasible to examine the exact indication for reoperation, thereby limiting our insights into potential reasons for TAA failure. However, we attempted to mitigate this limitation by only selecting codes for TAA revision, implant removal, and implant repositioning, rather than codes directly for irrigation, for our primary outcome of reoperation. Furthermore, we were likewise unable to examine the exact mechanism for postoperative periprosthetic fracture, allowing room for future research on this topic. Another limitation of our study is that we did not examine intraoperative periprosthetic fractures due to methodology limitations, although these fractures may be less common than postoperative periprosthetic fractures[28]. A key limitation of this study is that while our power analysis was designed for the primary outcome (reoperation), the secondary outcome of periprosthetic fracture may have been underpowered. As a result, the lack of statistical significance should be interpreted with caution, and larger studies are needed to better assess this potential association. Finally, as our data derived from US academic medical centers, it remains possible that outcomes are not generalizable to non-academic settings, or countries outside of the US which may have a different prevalence of osteoporosis or different surgical approaches for TAA.

While there was no increased risk of revision TAA or postoperative periprosthetic fracture in patients with osteoporosis in the present study, further research is warranted to corroborate our findings. Studies would ideally have a similar or larger sample size, include more granular measures of bone mineral density such as T-score, and include at least 3 years’ follow-up. In addition, the protective effects of concomitant prophylactic fixation strategies and implant design geared towards osteoporotic bone remain to be studied.