Knee osteoarthritis (KOA) is the most prevalent form of osteoarthritis, affecting 22.9% of people aged 40 years and over globally [1]. In fact, 45% of the population develops KOA in their lifetime [2]. The prevalence of KOA is expected to increase worldwide due to the increasingly aged and obese world population. The burden of KOA on individuals is significant due to pain and disability [3]. The burden of KOA on economies is also substantial, with 1 to 2.5% of the gross national product of countries such as the United States of America (USA), Canada, the United Kingdom, France, and Australia being spent to cover the health costs of osteoarthritis [4,5]. The largest part of the disease burden of KOA is attributed to total knee replacement (TKR), which is typically reserved for knees that progress to severe KOA.

There is no approved pharmacological treatment to slow, delay, or reverse the progression of KOA [6]. Given the high prevalence and significant individual and economic burden of KOA, new strategies to mitigate the progression of KOA are urgently needed. To approve such strategies, regulators require proof that these strategies result in the avoidance of or significant delay in structural endpoints, such as in the need for joint replacement [7]. However, the incidence of TKR is low: only 6 to 7% of the population aged 50 years and over will have TKR throughout the remainder of their lifetime, which is 6.4- to 7.5-fold lower than the above-mentioned 45% lifetime incidence of KOA [2,8]. This apparent discrepancy between the lifetime incidence of TKR versus KOA is related to the fact that TKR is influenced by multiple factors besides the progression of KOA, such as patient comorbidities, education, readiness for surgery, income, and health insurance [9,10]. The low incidence of TKR means that research measuring the incidence of TKR as an outcome requires an extended follow-up or a large number of participants to have sufficient cases that underwent TKR [11]. Indeed, to show a 50% reduction in the incidence of TKR over 5 years, with a power of 80% and 95% confidence, an interventional trial with two arms would require 362 to 5,459 participants, respectively, to complete the trial per arm (724 to 10,918 participants in total, respectively) [12]. The duration of follow-up or the number of participants required would be even greater for interventions that have modest effects on TKR, such as lifestyle modifications (e.g., diet or exercise and weight loss). For example, an observational study suggested that only a 20% reduction in the incidence of TKR over 8 years might be possible from a loss of 10% of initial body weight [13]. Therefore, instead of the incidence of TKR, other outcome measures [14] indicating severe KOA are essential to reduce the duration or sample size – and hence the cost – required for research into new strategies to mitigate the progression of KOA.

One such outcome measure has been developed and validated representing ‘end-stage knee osteoarthritis” (esKOA) [15]. An esKOA was defined if a knee meets one of two criteria: 1) moderate, intense, or severe symptoms of KOA indicated by pain and disability measurement and severe KOA indicated by radiographically-assessed knee structure; 2) intense or severe symptoms of KOA indicated by pain and disability measurement and frequent knee pain with mild or moderate KOA as indicated by radiographically-assessed knee structure. The esKOA is not a proxy for the outcome of incidence of TKR but rather an outcome to define the advanced stage of KOA, which is not influenced by the extraneous factors that contribute to TKR. It was demonstrated that the incidence rate of esKOA was more than three times higher than the incidence rate of TKR over 5 years (10.2% versus 3.0%, respectively) [15].

In the current study, we evaluated whether our esKOA definition [15] is responsive to an exposure known to have a modest association with the incidence of TKR (i.e., weight loss [13]). We hypothesized that our esKOA measure could reveal an earlier association with weight loss and require a smaller research cohort than TKR in both interventional and observational study designs. Confirming this hypothesis would suggest that adopting our esKOA measure as an outcome for the advanced stage of KOA could simplify and accelerate research, requiring fewer participants and yielding more efficient outcomes compared to using TKR as an outcome measure.