The health-economic model was developed from the perspective of healthcare payors to assess the financial impact on healthcare systems across European countries. The analysis focused primarily on the reimbursement burden borne by public healthcare systems, which represent the predominant funding mechanism in Europe. Reimbursement from public payors was assumed to approximate basic payments from the private sector, as the data obtained from the Statistical Office of the European Union (Eurostat) do not differentiate between public and private healthcare provision. This assumption can be considered conservative, since private healthcare delivery is generally associated with higher costs and reimbursement levels.

The model followed a five-step calculation process (Fig. 1; Alt et al., 2025). First, Eurostat data from 2023 on the number of primary hip and knee arthroplasties were identified; if not available for 2023, data from previous years were utilized (Fig. 1, Eq. 1). Second, primary implantations were combined with published infection rates to estimate the total number of hip and knee periprosthetic joint infections (PJIs) and combined with published rates for aseptic revisions to identify the number of primary aseptic revisions across 30 European countries (Fig. 1, Eq. 2; Springer et al., 2017; Straub et al., 2024; Szymski et al., 2023). Third, the rate of septic revisions following PJI treatment and following initial aseptic revision surgery was estimated based on previously published rates (Fig. 1, Eq. 3; Becker et al., 2025; Resl et al., 2024; Lenguerrand et al., 2022; Leta et al., 2015). Fourth, the re-revision procedures performed were estimated based on published treatment ratios for septic re-revisions (Lenguerrand et al., 2022). Revision procedures were classified into three categories: (1) debridement, antibiotics, and implant retention (DAIR); (2) one-stage exchange; and (3) two-stage revision procedures (Fig. 1, Eq. 4). Fifth, these procedure volumes were multiplied by the corresponding reimbursement payments (Fig. 1, Eq. 5). Reimbursement data were obtained through a survey conducted in 13 countries (Austria, Croatia, France, Germany, Italy, Lithuania, the Netherlands, Norway, Portugal, Slovenia, Switzerland, Türkiye, and the UK) and extrapolated to the remaining countries using mean gross domestic product (GDP) values as reported in Alt et al. (2025). Eurostat data were the “Surgical procedures” dataset (hlth_co_proc3, extraction ID 19281055). The Lenguerrand et al. treatment strategy distribution (DAIR 3.8 %; one-stage 23.1 %; two-stage 70.4 %) was derived from a single English/Welsh registry and applied uniformly across countries; because practice mix varies across Europe, reported costs represent a uniform-practice projection rather than country-specific observed costs.

Based on Alt et al. (2025), in the subsequent step, information on reimbursement payments received by hospitals from public healthcare funders for DAIR, one-stage, and two-stage PJI treatments was collected from 13 European countries through a survey coordinated by the EBJIS Country Delegates Group. For countries without direct data, a classification into higher- and lower-income groups was performed based on GDP per capita, sourced from Eurostat. GDP per capita ranged from 42 Purchasing Power Standards (PPS) in North Macedonia to 261 PPS in Luxembourg, with a European average of 101 PPS serving as the threshold for categorization (Table 1). Reimbursement amounts for each procedure type were multiplied by the corresponding number of interventions per country to estimate total payments per category (Fig. 1, Eq. 5). These were aggregated to national and European totals. Non-EUR costs were converted using 26 May 2023 exchange rates; public/private reimbursement equivalence was assumed for a conservative estimate. The GDP-tier extrapolation was validated by leave-one-out cross-validation (see Supplement).

A deterministic one-way sensitivity analysis was performed by varying key parameters by ±20 %: (1) incidence of primary PJI, (2) septic failure rate after PJI treatment, and (3) proportion of two-stage exchange procedures. Changes in PJI incidence and failure rates affected only the number of septic re-revisions after PJI, while infections after aseptic revision remained constant. For treatment strategy distribution, the two-stage share was modified by ±20 % relative to baseline, with the remaining proportion reallocated to one-stage procedures while keeping the total number of re-revisions unchanged. Total European reimbursement costs were recalculated for each scenario and compared with the base case.

To characterize compound parameter uncertainty, a probabilistic sensitivity analysis (10 000 Monte Carlo iterations; Beta-distributed rates, Dirichlet-distributed treatment mix, lognormal reimbursement values) was additionally performed; full distributional assumptions are detailed in the Supplement. The width of the resulting credibility interval is inherited directly from the ±20 % relative uncertainty assumed for the Beta- and Dirichlet-calibrated epidemiological and treatment-mix parameters, except for reimbursement values in GDP-tier-extrapolated countries, where the lognormal multiplier instead uses the empirical leave-one-out cross-validation error (70.7 % THA, 80.8 % TKA) rather than the ±20 % assumption.

According to the data provided by Eurostat, the Statistical Office of the European Union for 2023, there were 1 003 002 primary TKA and 1 284 887 primary THA performed, with a combined total of 2 287 889 primary THA and TKA procedures in Europe. Most procedures were performed in Germany (THA: 292 143, TKA: 223 133, THA + TKA: 515 276), followed by France (THA: 184 264, TKA: 140 047, THA + TKA: 324 311), the UK (THA: 124 322, TKA: 98 651, THA + TKA: 222 973), and Italy (THA: 130 640, TKA: 101 274, THA + TKA: 231 914) (Tables 2 and 3).

Assuming infection rates for TKA and THA of 1.03 % and 0.97 %, respectively, as reported by Springer et al. (2017) in a review analysing PJI incidence rates from several international arthroplasty registries, the estimated total annual number of hip and knee PJIs is 22 794. This figure includes 12 463 hip PJI cases and 10 331 knee PJI cases, arising from primary THA and TKA performed in 2023 (Tables 2 and 3). These base-case rates already represent an unweighted average across six national/regional arthroplasty registries (AOANJRR, New Zealand, Swedish Hip and Knee Arthroplasty Registers, National Joint Registry, American Joint Replacement Registry) rather than a single-source estimate. A more recent registry-based estimate combining the Danish Knee Arthroplasty Registry with the Danish national infection surveillance system reported a cumulative 1-year PJI incidence of 0.8 %–1.1 % for TKA, closely consistent with the 1.03 % base-case value used here; this comparison is retained as a validation check rather than a base-case change, as fully registry-stratified rates by joint and fixation type (cemented vs cementless) were not available for all 30 countries at the time of this analysis. To quantify the impact of this range on the model rather than reporting it only as a qualitative validation check, we re-ran the TKA reimbursement calculation with the PJI incidence fixed at 0.8 % and at 1.1 %, holding all other parameters at their base-case values. TKA re-PJI reimbursement costs, EUR 41,299,232 at the 1.03 % base case, changed to EUR 32 077 073 (-22.3 %) at 0.8 % and EUR 44 105 976 (+6.8 %) at 1.1 %; propagated through the full model, the European grand total shifted from EUR 123.6 million to EUR 114.3 million (-7.5 %) and EUR 126.4 million (+2.3 %) at the low and high ends of the Danish range, respectively. As this sensitivity range lies within the ±20 % deterministic bounds already reported above, we retain the Springer et al. (2017) base-case rate but present this arm explicitly rather than as a narrative aside.

Based on the reported septic re-revision rate after already performed surgical treatment of PJI of 30 % in THA and 21.7 % in TKA by Resl et al. (2024) and Becker et al. (2025), the number of septic re-revisions was 5981 (THA: 3739; TKA: 2242).

Simultaneously, to address septic re-revisions after an initial aseptic revision surgery in arthroplasties, a total of 23 882 aseptic revisions (THA: 12 849; TKA: 11 033) were identified, based on previous reported rates in the literature by Szymski et al. (2023) and Straub et al. (2024). Within the aseptic revision population, a total number of septic re-revisions of 648 (THA: 203; TKA: 445) was assumed based on Lenguerrand et al. (2022) and Leta et al. (2015) (Tables 2 and 3).

Based on the data of Lenguerrand et al. (2022), performed procedures in Europe were estimated for re-revisions, with 3.8 % being treated with a DAIR procedure, 23.1 % with a one-stage revision, and 70.4 % with a two-stage procedure.

By treatment strategy, THA re-revisions comprised 150 DAIR, 911 one-stage, and 2775 two-stage procedures; TKA comprised 102 DAIR, 621 one-stage, and 1891 two-stage procedures.

For THA, this comprised 142/864/2632 DAIR/one-stage/two-stage cases for re-PJI and 8/47/143 after aseptic revision.

For TKA, this comprised 85/518/1578 for re-PJI and 17/103/313 after aseptic revision.

Reimbursement per procedure was taken from published country-specific billing data for 13 countries, as reported by Alt et al. (2025), stratified by surgical strategy and joint (Table 1).

For re-PJI treatment in hip arthroplasty performed in 2023, total reimbursement payments across all European countries were estimated at EUR 70 257 309, with the highest country-specific costs of EUR 20 692 501 in Germany. For septic revision after initial aseptic revision surgery, the total costs of EUR 3 814 947 in Europe in 2023 were identified.

Major drivers were two-stage revisions, with a total of EUR 58 224 587 in re-PJI of the hip and EUR 3 814 947 in septic revisions after aseptic revisions (Table 4).

In knee arthroplasty, total reimbursement payments for re-PJI treatments were EUR 41 299 231, with the highest country-specific costs in Germany (EUR 12 260 418) and France (EUR 5 290 410). Septic revisions after an initial aseptic revision of TKA accounted for EUR 8 190 765 in Europe in 2023. Simultaneously, two-stage revision was the major cost driver in TKA for re-PJI and septic revisions after initial aseptic exchange, with costs of EUR 34 167 010 and EUR 6 776 247, respectively (Table 5).

Overall, European healthcare systems are expected to face a total reimbursement burden of EUR 123 562 255 (≈ EUR 123.6 million) for septic re-revision surgery performed in 2023. Because the two source populations are epidemiologically distinct (see Methods), this total is reported as the sum of two separate estimates rather than a single aggregated figure: EUR 111 556 541 for re-revision following surgical treatment of primary PJI (THA EUR 70 257 309; TKA EUR 41 299 232) and EUR 12 005 714 for septic re-revision after an initially aseptic revision (THA EUR 3 814 948; TKA EUR 8 190 766). Of this total, the 13 directly surveyed countries (Austria, Croatia, France, Germany, Italy, Lithuania, the Netherlands, Norway, Portugal, Slovenia, Switzerland, Türkiye, and the UK) account for EUR 98 547 420 (79.8 %), with the GDP-tier extrapolation for the remaining 21–22 countries contributing the remaining EUR 25 014 835 (20.2 %); this subtotal is reported alongside the total, rather than in place of it, so that readers can judge the relative weight of directly observed versus extrapolated reimbursement data.

Varying PJI incidence or septic failure rates by ±20 % changed overall costs to EUR 101.3–145.9 million (±18.1 %); varying the two-stage share by ±20 % changed costs to EUR 114.5–132.6 million (±7.3 %). Fluctuations were larger for THA (EUR 68.6–79.6 million) than TKA (EUR 46.0–53.0 million), reflecting THA's greater share of baseline expenditure.

A country-level Monte Carlo PSA (10 000 iterations; Methods) yielded a median total of EUR 126.2 million (95 % credibility interval EUR 102.0–156.2 million; THA EUR 55.5–101.3 million, TKA EUR 38.2–66.6 million), consistent with the deterministic bounds above.

Leave-one-out cross-validation of the GDP-tier extrapolation showed a mean absolute percentage error of 70.7 % (THA) and 80.8 % (TKA), confirming GDP per capita is an imperfect proxy for individual-country reimbursement, although errors partially offset when aggregated to the European total (see Supplement).

Two counterfactual scenarios were modelled to illustrate the potential impact of prevention and treatment-pathway optimization. A 25 % reduction in PJI incidence (holding all other parameters at base-case values) reduced the projected total to EUR 95.7 million, a saving of EUR 27.9 million per year (22.6 %), avoiding an estimated 935 THA and 560 TKA re-revisions. Shifting 50 % of two-stage candidates to one-stage exchange (holding incidence constant) reduced the projected total to EUR 100.9 million, a saving of EUR 22.6 million (18.3 %). Combining both interventions reduced the projected total to EUR 78.2 million, a saving of EUR 45.4 million (36.8 %).