. 2020 Oct 8;22:485–492. doi: 10.1016/j.jor.2020.10.003

Table 1.

Summary of the ICERs and key findings of all 33 CEA studies included in this review, stratified by comparator type.

Author (year, country)	Study Design	ICER^a	WTP^a	Key findings
UKA vs TKA
Soohoo et al. (2006, USA)³³	Decision Tree	Implant survival:	$50,000	UKA is a cost-effective alternative for unicompartmental OA when the durability and function of the UKA implant is similar to the TKA one and with appropriate patient selection. However, when adjusted for 2020, UKA is no longer cost-effective to TKA for the 17 y vs 20 y category.
		12 y vs 15 y: $440,
		11 y vs 15 y: TKA dominant, 17 y vs 20 y: $59,556^{^}
Slover et al. (2006, USA)⁵¹	Markov	Base case: UKA dominant	$50,000	Although UKA was dominant, the average differences in costs and QALYs with TKA were $200 and 0.05 respectively. Thus, both have similar cost-effectiveness in the elderly low-demand population.
Peersman et al. (2014, Belgium)⁵²	Markov	UKA dominant	€10,000, €25,000, €50,000	UKA offers a clear increase in health outcomes for a smaller cost than TKA for unicompartmental OA.
Ghomrawi et al. (2015, USA)³⁴	Markov	Age at surgery (TKA vs. UKA):	$100,000	UKA is cost effective compared to TKA in patients over 65 years of age for end-stage unicompartmental OA. However, modest improvements in implant survival could make it cost-effective in younger patients.
		45 y: $34,011
		55 y: $71,425
		65 y,75 y,85 y: UKA Dominant
Chawla et al. (2017, USA)³⁵	Markov	PFA vs TKA:	$50,000	Improvements in implant survival have resulted in PFA being a more cost-effective joint preserving procedure in younger patients.
		50 y - $3445
		60 y - $ 3405
Burn et al. (2018, UK)⁵³	Markov	UKA Dominant	£20,000	UKA is expected to provide better health outcomes and lower lifetime costs than TKR. However, surgeon usage of UKA has a significant impact on cost-effectiveness.
Xie et al. (2010, Canada)⁵⁴	Prospective cohort	Perspective (TKA vs. UKA):	$10,000 and $50,000	There is a 0.4 probability of TKA being cost-effective from the societal or patient's perspective at a $50,000 WTP and 0.7 from the government's at a WTP of $10,000 for unicompartmental OA (without inflation adjustment). Longer study needed.
		Societal - $79,430
		Patient - $73,510
		Government - $5917
Beard et al. (2019, UK)⁵⁵	RCT	PKR dominant	n/a	During the 5-year study period, PKR offers slightly better outcomes, lower surgical costs and lower follow-up health-care costs compared to TKR for treatment of late-stage isolated medial compartment OA. Hence, it should be first choice for it.
Resurfacing vs No Resurfacing in TKA
Weeks et al. (2018, Canada)³⁶	Decision Tree	Resurfacing Dominant	n/a	Resurfacing the patella is cost-effective due to higher revision rates for non-resurfaced TKA.
Zmistowski et al. (2019, USA)³⁷	Decision Tree	Overall: $3,032^b,	n/a	It is not cost effective to routinely resurface nonarthritic patella during primary TKA. Selective resurfacing for arthritic patients is vital for cost-effectiveness.
Zmistowski et al. (2019, USA)³⁷	Decision Tree	Nonarthritic patellae only: $183,584^b	n/a
Computer-Assisted TKA vs Non-Assisted TKA
Novak et al. (2007, USA)³¹	Decision Tree	$59,033^{^}	$50,000/	Computer-assisted implant alignment systems increase the precision of component alignment enough to reduce failure rates and revisions to justify the extra cost vs. mechanical alignment systems during TKA (not true following 2020 ICER adjustment).
Novak et al. (2007, USA)³¹	Decision Tree	$59,033^{^}	$100,000
Dong et al. (2006, UK)³⁰	Markov	CAS dominant vs. Conventional TKA	£30,000	Computer-assisted TKA is cost effective in the long-term through reducing revision rates and complications via more precise component alignment.
Gothesen et al. (2013, Norway)²⁹	Markov	Conventional TKA dominant vs CAS	500,000 kr (NOK)	At high operation volume hospitals, CAS needs to improve implant survival in 60 and 75-year-old cohorts just marginally for cost-effectiveness. A more significant increase is needed for low volume hospitals.
TKA vs Non-Operative Management
Losina et al. (2009, USA)³⁹	Markov	Overall: $21,123	$50,000/$100,000	TKA is cost-effective across all risk groups for perioperative complications.
Losina et al. (2009, USA)³⁹	Markov	High-risk patients only: $36.415	$50,000/$100,000
Ponnusamy et al. (2018, Canada)⁴⁰	Markov	Non-obese: $3510, Overweight: $3,002, Obese: $3,118, Severely obese: $3,742, Morbidly obese: $5,853, Super obese: $12,569	$50,000	Not opting for TKA care based on BMI is not justified, even in the super obese cohort.
Elmallah et al. (2017, USA)⁵⁶	Prospective	$47,357	$50,000	Following an OA diagnosis, TKA is cost-effective.
Dakin et al. (2012, UK)³⁸	RCT	Baseline patient OKS:	£20,000 -£30,000	The UK primary trust criteria (2012) restricting TKA to patients with pre-operative OKS <27 is denying a cost-effective treatment to patients above this OKS.
		<9: £5,768,
		9-11: £5,577,
		12-13: £5,032,
		14-15: £5,152,
		16-17: £6,407,
		18-19: £8,068,
		20-21: £14,735,
		22-24: £11,270,
		25-27: £13,655,
		>27: £14,366,
Skou et al. (2020, Denmark)⁵⁷	RCT	TKR plus non-surgical treatment vs. non-surgical treatment only:	€22,655	From a 24-month perspective, in patients with moderate to-severe knee OA in secondary care in Denmark, TKR plus non-surgical (exercise, education, diet, insoles and pain medication).
Skou et al. (2020, Denmark)⁵⁷	RCT	Unadjusted base-case: €19,579, Adjusted base-case: €34,519, Adjusted base-case including deaths: €48,984 to 67,964	€22,655	treatment is not cost-effective compared with non-surgical treatment with the potential for later TKR if needed when adjusted for the covariates: age, sex and baseline values.
TKA Implant Type
Fennema et al. (2014, Germany)⁴⁶	Markov	TKA with advanced low-wear bearings vs standard polyethylene bearing: overall: €18,198, patients <55 y: €722 and 75: €91,687^b	€0, €10,000, €25,000, €50,000	Low-wear articulations may be considered cost-effective overall but it is age-dependent, with the ICER being significantly lower for younger people than for older people, where it no longer becomes cost-effective.
Pennington et al. (2016, UK)⁴⁸	Markov	AGC Biomet dominated Genesis 2 and Triathlon. PFC Sigma dominated by Nexgen.	£20,000 - £30,000	AGC Biomet prostheses are the least costly cemented unconstrained fixed brand for TKR but Nexgen prostheses lead to improved patient outcomes, at low additional cost and so should be first choice as they are the most cost-effective.
Pennington et al. (2016, UK)⁴⁸	Markov	For 70 y men and women, Nexgen vs. AGC Biomet, £2715 and £2667 respectively	£20,000 - £30,000
Suter et al. (2013,USA)⁴⁷	Markov	Innovative vs standard TKA implants: ≥ 50% decrease in long-term TKA failure at ≤ 50% increased cost: < $100,000.	$150,000	Innovative implants must decrease TKA failure by 50–55% or more compared to standard implants to be broadly cost-effective.
Suter et al. (2013,USA)⁴⁷	Markov	A 20% decrease in long-term failure at 50% increased cost: < $150,000 (only in healthy 50,59 y)	$150,000
Hamilton et al. (2013, UK)⁵⁸	RCT	Over a lifetime and at 1 year, Triathlon TKA dominated Kinemax TKA	£20,000	The values for money saved per QALY were statistically insignificant and so both implants were of similar value using the SF-6D and QALY methodology.
Multiple Surgical Comparator studies
Konopka et al. (2015, USA)⁵⁹	Markov	50-60 y:	$50,000/$100,000/$150,000	In a 50-60 y with unicompartmental medial knee OA, HTO is the most cost-effective management option. TKA is also more cost effective than UKA in the same cohort.
		TKA vs HTO - $262,908,
		TKA vs UKA - $14,058^c
Kazarian et al. (2018, USA)⁶⁰	Markov	TKA dominates NST from 40 to 69 years and just over $16,494 at 80 y, UKA dominates TKA for all ages at time of treatment	$50,000	In unicompartmental OA, using surgical treatments is cost-effective in all age groups. UKA should be prioritised over TKA.
Stan et al. (2015, Romania)⁴⁵	Prospective	TKA to unoperated knee dominates both TKA after HTO and CM	n/a	Careful patient selection could help optimise the cost-effectiveness of TKA as unoperated knee TKA is dominant to operated knee TKA.
Murray et al. (2014, UK)⁶¹	RCT	Patellar resurfacing vs. no resurfacing: >95% probability of being cost-effective at WTP,	£20,000	Patellar resurfacing is cost-effective, mobile bearings highly cost-effective (however there was considerable uncertainty), all-polyethylene components are poor value for money and should not be used in place of metal-backed components.
		Mobile bearing vs fixed bearing: £2044
		Metal-backed tibial components vs all-polyethylene ones: £43
Hak et al. (2013, UK)⁴⁴	n/a	Lifetime + 10-year durability:	£20,000–£30,000	In the NHS, the KineSpring System is the most cost-effective strategy to treat knee OA.
		UKA, HTO and KineSpring system dominated TKA, TKA dominated CM.
		TKA vs no treatment:
		£1,303^b (lifetime)
		€4,153^b (10 year)
Marcacci et al. (2013, Italy)⁴¹	n/a	Lifetime + 10 year durability:	€25,000–€30,000	In the Italian Healthcare system, the KineSpring system offers the lowest cost/QALY and so is the most cost effective option.
		UKA, HTO and KineSpring system dominated TKA, TKA dominated CM, TKA vs no treatment:
		€2348^b (lifetime)
		€4,884^b (10 year)
Li et al. (2013, Germany)⁴²	n/a	KineSpring System dominated surgical treatments (TKA, UKA and HTO) and conservative management.	€ 39,742	The KineSpring System is the most cost-effective option for knee OA patients in Germany.
Li et al. (2013, Germany)⁴²	n/a	Surgical treatments vs. no treatment: €10,722	€ 39,742
Strain et al. (2015, Spain)⁴³	n/a	Lifetime + 10-year durability:	€20,000 to €30,000	Same results achieved in the Spanish Healthcare system as Italian.⁴¹ The KineSpring System is the most cost-effective treatment for knee OA.
		UKA, HTO and KineSpring system dominated TKA, TKA dominated CM. TKA vs no treatment:
		€2530 (lifetime)
		€5264 (10 year)
Other:
Odum et al. (2013, USA)²⁰	Markov	Simultaneous bilateral TKA dominated staged bilateral TKA	$328,874	Simultaneous bilateral total knee arthroplasty is more cost-effective with lower costs and greater health outcomes for the average patient.
Van der Woude et al. (2016, Netherlands)⁶²	Markov	KJD dominant vs. TKA	€ 20,000	Treating knee OA with KJD over TKA has a high potential to be cost-effective, which is most likely in the younger population.
Clement et al. (2019, UK)⁶³	Markov	Robot-assisted UKA vs TKA by annual patient case volume (patients/year):	£20,000	Robot-assisted UKA is cost-effective compared with manual TKA for patients with isolated medial compartment knee OA. Increasing the annual patient case volume and reducing the length of hospital stay decreased the ICER of using rUKA over manual TKA.
		10: £7170^b,
		100: £1395^b,
		200 (with a 2 day stay): £648^b,
		200 (with a 1 day stay): £364^b

BMI – Body Mass Index, CAS – Computer assisted surgery, CM – Conservative management, HTO – High Tibial Osteotomy, KJD - Knee Joint distraction, NST – Nonsurgical treatment, OA – Osteoarthritis, OKS – Oxford Knee Score, PFA – Patellofemoral arthroplasty, PKR – Partial Knee replacement, QALY – Quality-adjusted life year, RCT – Randomised controlled trial, SF-6D – Short-Form 6-Dimension health index, UKA - Unicompartmental knee arthroplasty, WTP – Willingness to pay threshold, y = Years of age.

^{^} Inflation adjustment resulted in ICER increasing above WTP threshold and so the management option no longer being cost effective. This is reflected in the main findings being altered.

All values are given per QALY.

Inflation year not given in study, so 2020 adjustment not performed.

ICER value has been calculated using provided information in study.