Background The complex temporal relationship between the nature and timing of symptom onset/worsening and underlying structural progression lies at the heart of epidemiological studies of the natural history of symptomatic osteoarthritis (OA). In previous work we identified the existence of a prodromal phase of symptom worsening beginning on average 2-3 years prior to the first appearance of radiographic knee OA (Kellgren & Lawrence (KL) grade≥2). It remains unclear whether this upward trajectory of prodromal symptoms marks the beginning of longer term deterioration or whether they plateau or regress, as might be predicted by a phasic model of OA progression.
Objectives To investigate the trajectory of self-reported pain, stiffness, function and other symptoms following the incidence of radiographic knee OA.
Methods Data for these analyses are from the Osteoarthritis Initiative (OAI) public use data sets, which uniquely provides patient-reported outcome measures and plain radiography of the knees repeated annually. Knees which had developed new tibiofemoral radiographic OA (KL grade≥2) at any of the first 4 annual follow up visits, without frequent knee pain, aching or stiffness at enrolment, were defined as a case. Cases were assigned a common baseline time-point, t0, corresponding to the time of incident radiographic OA. Symptoms investigated were knee-specific WOMAC and KOOS pain, function, stiffness and other knee symptom subscale scores and their individual items (n=33), available up to 3 years before and up to 5 years after incidence of radiographic OA. Knees that were replaced were censored at the last visit before the knee replacement was recorded. Trajectories of the probabilities of having at least one of the symptoms on a subscale, and for each of the individual symptoms over time were fitted using multi-level logistic regression models, adjusting for age and gender and allowing for repeated measures within individuals.
Results 169 cases of incident radiographic knee OA were recorded in 161 participants. 22 (13%) of these progressed to a higher KL grade after incidence during the first 4 years of OAI follow-up and 6 (4%) had a total knee replacement. Following the initial prodromal phase, the probability of symptoms generally plateaued although each of the subscales showed subtly different trajectories (Figure, time 0 represents time of incident radiographic knee OA). The probability of moderate, severe or extreme symptoms was consistently low. An exception was frequency of experiencing pain which increased greatly in the lead up to incidence and then decreased slightly after incidence. The probability of grinding, hearing clicking or any other type of noise when the knee moves also had a higher probability, increasing before incidence, plateauing around incidence and then starting to increase again around 3 years after incidence.
Conclusions Our study found that following an increase in symptoms during the prodromal phase, symptoms generally do not keep getting worse after the incidence of radiographic knee OA. Instead, consistent with a phasic model of OA progression, in the absence of further structural progression they typically plateau or even reduce in the period up to 5 years after incidence.
Disclosure of Interest None declared