Registry Cohort Study · Knee Arthroplasty

Knee arthroplasty in patients under 50 years demonstrates durable survivorship despite substantially higher lifetime revision risk than in older patients

A nationwide registry analysis of the Scottish Arthroplasty Project

Morrison SR, Hall AJ, Lim B, Akhtar A, Walmsley P

Knee Surg Sports Traumatol Arthrosc 2026;1–11 · doi:10.1002/ksa.70486

3,069

Patients Aged 16–49 Years

15.5%

21-Year Cumulative Revision Incidence

31.3%

Estimated Lifetime Revision Risk

8 yrs

Median Follow-Up (Range 1–21)

As Published Knee Surgery, Sports Traumatology, Arthroscopy journal cover

European Society of Sports Traumatology, Knee Surgery & Arthroscopy (ESSKA)

This study was published in Knee Surgery, Sports Traumatology, Arthroscopy (KSSTA), the clinical journal of ESSKA. Using the Scottish Arthroplasty Project — a nationwide registry capturing all arthroplasty procedures performed in Scotland — it characterises the epidemiology, implant survivorship, and revision risk of primary knee arthroplasty in patients aged 16–49 years, with a date-matched cohort aged ≥50 years for comparison.

Journal KSSTA

Data Source Scottish Arthroplasty Project

Study Design Registry cohort

Level of Evidence III

Read on wiley.com View e-PDF

Abstract

Purpose

Data on survivorship and epidemiology in patients aged <50 years undergoing primary knee arthroplasty (KA) are limited. This study used nationwide registry data to evaluate epidemiology, implant survivorship, and revision risk in patients aged 16–49 years.

Methods

This retrospective analysis of prospectively collected data from the Scottish Arthroplasty Project evaluated patients aged 16–49 years undergoing primary KA between 2000 and 2019. A date-matched cohort aged ≥50 years served as comparison. Data were analysed using SPSS and R. Outcomes included revision rate, time to revision, and mortality. Competing risk analysis assessed cumulative incidence, and lifetime revision risk was estimated using parametric survival modelling.

Results

3069 patients (58.6% female; mean age 44, range 16–49) were included, with median follow-up of 8 years (range 1–21). Non-osteoarthritis indications were more common in patients aged <35 years. KA incidence increased significantly over time in patients aged <50 (IRR 1.05/year, p<0.001). Overall, 8.3% underwent revision, with a 21-year cumulative incidence of 15.5%. Patients aged ≥50 years had lower revision risk (HR 0.4, p<0.001). Estimated lifetime revision risk was 31.3% (95% CI 24.6–38.4), with wide confidence intervals in younger subgroups. Mortality at 21 years was 18.2%, higher in nonosteoarthritis indications (HR 0.5 for OA, p<0.001).

Conclusion

KA in patients under 50 years demonstrates acceptable long-term survivorship but a higher revision risk compared to older populations, with an estimated lifetime revision risk of approximately 30%. Subgroup findings should be interpreted cautiously due to small sample sizes and key confounders, particularly implant type and underlying diagnosis. These results support KA in selected younger patients while highlighting the importance of counselling regarding long-term revision risk and the need for more granular future research.

Level of Evidence

Level III.

View paper on wiley.com

Cohort & Design

Study cohort

SAP

All primary KA in Scotland, 2000–2019

→

3,069

Patients aged 16–49 years

→

8 yrs

Median follow-up (range 1–21)

→

256

Underwent revision (8.3%)

A date-matched cohort of 117,319 patients aged ≥50 years served as comparison. Data were linked to each patient's Community Health Index number, enabling accurate nationwide follow-up for revision, complications, and mortality. Analyses used cumulative incidence functions with the competing risk of death, Cox proportional hazards regression, and parametric (Weibull) survival modelling for lifetime revision risk.

Findings

Six key findings

Rising Utilisationn=3,069

KA in patients <50 accounted for 2.3% of all primary KA in Scotland, and incidence rose significantly over the study period (IRR 1.05/year, p<0.001) — an approximately 2.6-fold (164.7%) increase between 2000 and 2019.

Osteoarthritis was the indication in 79.7% of cases; non-OA indications were more common in younger patients, rising to a majority of cases in those aged <35 years.

Most procedures were unilateral (88.7%) and used cemented implants (87.3%).

Implant Survivorship21-year

Overall, 8.3% (256/3069) of patients underwent revision, with a median time to revision of 12 years (range 1–21).

The cumulative incidence of revision at 255 months (21.3 years), accounting for the competing risk of death, was 15.5% (95% CI 13.0–18.0).

This represents acceptable, durable long-term survivorship in a demanding younger population.

Revision Risk vs Older PatientsHR 0.4

Patients aged ≥50 years were 60% less likely to undergo revision than the <50 cohort (HR 0.4, 95% CI 0.36–0.46, p<0.001).

Revision rate peaked at 9.6% in the 40–44 age group; the youngest patients (16–25) had the lowest observed rate at 3.1%.

Within the <50 cohort, age itself was not an independent predictor of revision (p=0.089), suggesting a more complex relationship than a simple linear age effect.

Lifetime Revision Risk31.3%

The estimated lifetime risk of revision for the whole 16–49 cohort was 31.3% (95% CI 24.6–38.4), modelled across the remaining lifetime.

Subgroup estimates ranged from 19.3% (25–29 years) to 35.3% (40–44 years), but were associated with wide confidence intervals.

These are modelled projections and should be considered exploratory, particularly in the smallest age groups.

Revision Indicationsn=256

Aseptic loosening was the dominant mode of failure — 162/256 revisions (63.6% of all; 91.5% of those with a recorded indication).

Infection accounted for 15/256 revisions (5.9% of all; 8.5% of recorded), broadly comparable with wider arthroplasty populations.

Revision indication was unavailable in 30.5% of cases, so these proportions are reported as bounded ranges rather than definitive.

Mortality18.2%

Raw mortality in the <50 cohort was 6.2% (190/3069); the cumulative incidence of death at 21.3 years was 18.2%, approaching that of revision over long-term follow-up.

Mortality was significantly higher for non-OA indications (HR 0.5 for OA, p<0.001), and male sex independently predicted mortality (HR 1.4, p=0.040).

Overall mortality exceeded the general population rate for this age group (6.2% vs 4.6%), likely reflecting the burden of non-OA disease.

Lifetime Revision Risk by Age

Estimated lifetime risk of revision, by age at primary arthroplasty

Age at surgery (years)	Individuals (n)	Revisions (n)	Lifetime risk % (95% CI)
<25	32	1	25.3 (2.9–78.3)
25–29	60	2	19.3 (4.2–50.9)
30–34	107	7	27.8 (13.5–48.9)
35–39	203	14	28.5 (16.9–43.7)
40–44	675	65	35.3 (26.3–45.5)
45–49	1,992	167	30.9 (24.4–38.2)
Total	3,069	256	31.3

Table 3 from Morrison et al. (2026), Knee Surgery, Sports Traumatology, Arthroscopy. Lifetime risk estimated using single-decrement life-table methodology; wide confidence intervals in the smallest subgroups reflect low event numbers and should be interpreted with caution.

Limitations & Future Research

Interpreting the findings, and where the evidence must go next

Total versus unicompartmental KA could not be distinguished

The registry did not separately code total and unicompartmental knee arthroplasty for most of the period. As implant type is strongly associated with revision risk, differences in implant use across age groups represent a significant unmeasured confounder. Future datasets should differentiate implant types.

Subgroups were small and underpowered

Several strata, particularly the youngest patients, contained small numbers and few revision events, producing wide confidence intervals. These analyses are hypothesis-generating rather than definitive; larger collaborative or multi-registry studies are needed for precise estimates in the very young.

Missing data limit inference on revision indication

Revision indication was unavailable in 30.5% of revised cases, and the dataset does not code secondary patellar resurfacing as a revision — potentially underestimating revision rates and limiting comparability with other registries.

Patient-level variables and functional outcomes are needed

The absence of comorbidity, activity level, socioeconomic data and patient-reported outcome measures limits confounder adjustment and assessment of functional benefit. Future research should incorporate granular patient-level variables, improved non-OA classification, and integration of PROMs alongside revision risk.

Take-Home Messages

Primary knee arthroplasty in patients under 50 demonstrates acceptable long-term survivorship (15.5% cumulative revision incidence at 21 years), but carries a substantially higher revision risk than matched patients aged ≥50 (HR 0.4 in the older group).

The estimated lifetime revision risk of approximately 31% underscores the importance of counselling younger patients about long-term revision burden when planning surgery.

Subgroup findings are hypothesis-generating given small samples and key confounders — particularly implant type and underlying diagnosis — and more granular, multi-registry research is needed to refine outcome estimates in this population.

Knee arthroplasty in patients under 50 years demonstrates durable survivorship despite substantially higher lifetime revision risk than in older patients

Study cohort

Six key findings

The cohort at a glance

Estimated lifetime risk of revision, by age at primary arthroplasty

Interpreting the findings, and where the evidence must go next