**4. Discussion**

Despite a rising number of projection studies for hip and knee arthroplasties in recent literature, we still lack empirical results on studies for the upper extremities. Although current growth rates for shoulder arthroplasties are reported to be comparable with, or even higher than the growth rates for total hip and knee procedures [8,10], the demand and impact of shoulder arthroplasty have been underestimated in the past. However, based on evolving scientific evidence regarding good functional outcomes and due to an expanding indication spectrum, shoulder arthroplasty has been one of the main focus in orthopedic surgery in recent years. Therefore, we opted for a data-driven model selection, by minimization of model errors to investigate the trends for primary and revision shoulder arthroplasty procedures from 2010 to 2040. Although both models predicted a slightly different total number due to their distinct intrinsic model assumptions, they showed an important trend during the upcoming decades. Using these models, the number of total shoulder replacements was estimated to grow up to 700%, mainly due to rising adoption of rTSA in fracture-related conditions of the elderly. These growth rates are substantially higher than the current rates of hip and knee arthroplasty [2,7], indicating the immense importance for orthopedic surgeons in the future.

With the total number and incidence of primary TSA procedures rising, we also modeled a rising number of revision procedures, highlighting its increasing burden in the future.

#### *4.1. Projections of Shoulder Arthroplasty and International Comparison*

When comparing our projections with other countries, we found that the trend of a rising volume of primary shoulder arthroplasties has also been reported in the United States [10,23,24]. Schwartz et al. detected an over five-fold increase in primary shoulder arthroplasty between 2001 and 2010 [24], while Wagner et al. projected even higher growth rates in upper extremity arthroplasty compared with total hip and knee procedures by the year 2025 (+235% in total volume) [25]. Padegimas et al. even projected the future volume to increase by 755% by 2030. Similar to our results, this can mainly be attributed to an exponential rise in the number of reverse shoulder arthroplasty in the upcoming years, with its prevalence already rising from 7.3 to 19.1 per 100,000 in the U.S. However, the demographic and economic characteristics of the United States cannot easily be transferred to other industrialized countries, although the global demographic pattern of an aging population is identical around the world. However, projections like those of Wagner et al. [25] tend to be overestimated due to a steep increase of procedure volume in the baseline years- a problem we also faced using a Poisson model. For this reason, we implemented an alternative methodical approach, which included a rather asymptotic development of shoulder arthroplasty in the future but still estimated an increase in incidence that is several times higher than the projected increase in hip or knee arthroplasty [2,4,5,26].

These findings are also in line with the most national recent trends for shoulder arthroplasty in Europe. Although population and economic data, as well as procedure selection for the major indications, vary between most countries, Lübbeke et al. [13] reported a strong upward trajectory in the incidence of shoulder arthroplasty in nine registries, concluding that due to growing demand, increasing health care capacity, and/or expanding indications, countries could expect to see this continuation in growth in the future. According to the most recent European study of Villate et al. derived from the French hospital discharge database, the number of total shoulder arthroplasties is projected to rise up to 322% by 2050, representing a future challenge for their healthcare system [27].

#### *4.2. Projections as a Function of Age and Indication*

Based on our data projections, this trend is mainly attributed to a considerably increasing utilization and incidence of rTSA in patients older than 80 years, although the number and incidence of procedures performed for younger patients are forecast to rise in the future as well.

While this observed trend is in line with the findings from other countries [13,28], including the United States [9], it is in contrast with the trend seen in hip and knee arthroplasties, where the incidence rate in younger and older patients remained constant over all age groups [2]. This difference may be explained by various factors. First, based on the recent population growth in Germany, the share of patients > 65 years of age is projected to increase from 21% to 30% by the year 2040, which may contribute to a rapid increase in rTSA utilization, as reverse prostheses are rarely implanted in younger patients. Second, the indications for rTSA have been expanding during the last decade, as indicated by the almost ten-fold increase in the number of proximal humerus fractures in this study. Many surgeons may also now see it as a possible solution for poor shoulder conditions, which were previously seen as unsolvable [29]. In complex proximal humerus fractures, which represent the third-most common fracture seen in patients aged > 60 years [30], rTSA was found to provide more reproducible function with better recovery, as well as lower revision rates than HA [31,32]. However, recent studies also suggest, that shoulder arthroplasty due to fracture could be associated with more inconsistent outcomes [33,34] and an increased risk of postoperative complications compared with OA and cuff arthropathy [35]. Additionally, shoulder arthroplasty for fracture seems to lead to an even higher resource utilization [36], which is especially important as we documented a massive increase in these procedures in the upcoming decades, probably affecting future healthcare costs. Against the background of moving toward the era of bundled paymen<sup>t</sup> models, an appropriate risk adjustment based on the indication of surgery should, therefore, be promoted to maintain the current standard of care for all patients.

Although we documented a massive increase in fracture-related rTSA in our study, current scientific evidence contrarily suggests, that non-operative treatment may be favorable in the elderly compared to surgical treatment in certain fracture patterns [37,38]. However, based on the increasing demand for self-independency in the elderly, there is still a controversial debate, on which patients may be better treated by surgical intervention. Therefore, it remains to be seen, if this trend may influence the treatment choice and amount of implantations in the future, as the number of proximal humerus fractures will be constantly rising in many European countries over the next decade [39,40].

#### *4.3. Projections of Revision Shoulder Arthroplasty*

With the total number and incidence of primary TSA procedures rising, we also modeled a rising number of revision procedures, which has globally never been performed up to date. While we were not able to calculate survival or revision rates because of the aforementioned issue, we used the RB as a surrogate parameter to assess the procedurespecific risk for revision. As a consequence, the actual revision risk, especially for TSA, appears to be underestimated by our calculations, which is why we also used a constant rate approach to estimate the number of revision surgeries based on the RB during the baseline years, which still led to an up to 8.4-fold increase by 2040.

Based on our models, revision surgery for TSA will be significant in 2040 and would be expected to increase even more over time as a direct result of the maturation of the increased number of primary surgeries being performed. This raises the concern if there will be enough resources and trained surgeons to carry out these difficult revision procedures in the future. Against the background of a rising number of fracture-related rTSA, this also highlights the enormous need for adequate fracture-prevention programs, in order to limit the number of revision surgeries in the future, as well.
