The overall prevalence of foot pain in people with symptoms of knee or hip OA attending the GLA:D® programme was 12%. The number of other painful knee/hip joints and pain severity in worst knee/hip joint, were associated with foot pain for those with knee or hip OA. Use of pain medication was associated with foot pain in those with index knee OA only. No associations were seen for physical activity level and foot pain in either those with index knee or hip pain.

The prevalence of foot pain was slightly lower than previously reported general population prevalence estimates of 13–36% [4]. The definition of foot pain in GLA:D® was based on pain during the last 24h, which does not reflect fluctuations in foot pain over a longer period of time, nor does it consider symptoms such as aching or stiffness unlike other cohort studies [4]. Paterson et al. [5] previously reported a notably higher prevalence of foot pain (25%) than identified in this work, in their cohort analysis of people with symptomatic radiographic knee OA, defined by frequent knee symptoms (including pain, aching, or stiffness in and around the knee on most days of the month for at least one month in the past year) and radiographic evidence of knee OA (Kellgren/Lawrence grade ≥ 2). The difference could be accounted for by variance in sampling strategy between studies, and thus the cohort reported within this study may represent a wider spectrum of disease, including those with potentially earlier pathogenesis and minimal radiographically observable change. This is the first study to determine a prevalence estimate of foot pain in those with symptomatic hip OA and as such similar comparisons cannot yet be drawn.

It is unclear from cross-sectional observation alone whether foot pain precedes, coincides with, or follows the presence of symptomatic knee or hip OA. Clinically, such information could provide useful guidance for future treatment targets and as such future longitudinal studies are recommended. For example, identification of people for whom symptomatic knee OA is likely to precede concomitant foot pain could lead to earlier treatment targeting to minimise foot pain. This study has however highlighted that whilst the direction of this relationship is unclear, the odds of having foot pain is significantly increased with knee or hip pain severity and the presence of pain in more than one hip or knee joint. This highlights the importance of including foot pain measures within the clinical assessment of knee and hip symptoms, particularly in those with a diagnosis of OA. Given that it is known that people with pre-operative foot pain are more likely to have poorer clinically important outcomes following knee arthroplasty [18], the potential for foot pain to negatively impact concurrent knee and hip symptoms and vice versa should therefore be considered.

Participants of the GLA:D® programme with knee or hip OA symptoms and foot pain were more likely to experience worse pain intensity and pain in more sites. Furthermore, those with symptomatic knee OA and foot pain were also more likely to be using analgesic medicine. The findings reported here are consistent with similar prior observations, where worsening knee pain was associated with foot symptoms over four years in those with symptomatic radiographic knee OA [6]. It is noteworthy that factors such as worsening pain have also been associated with poorer outcome or disease progression in previous studies of knee and hip OA interventions [19, 20]. Arguably, based on our findings, future consideration should also be given to the effect that foot health/pain could have upon the likely success of various treatments for knee or hip OA, including exercise and surgical intervention. For example, it is currently unclear to what extent exercise therapy benefits or is limited by foot pain. Future epidemiological studies could seek to build upon the data presented here to explore the wider impact of foot pain upon people living with symptomatic knee or hip OA. We may then consider the need to optimize foot health as a potentially modifiable treatment target for wider determinants of lower limb health or health-related outcomes.

Some of the limitations of this study, such as use of this particular pain question and cross-sectional design, have already been described. In addition, there is a potential lack of generalisability as the current study only includes those who sought education and exercise for their knee/hip OA pain, whilst other factors that were not accounted for (e.g., acute injury to the foot, previous injury or surgery to the foot, footwear, occupation, etc.) may have contributed to the associations found. We do not know the reasons for foot pain, nor the longevity of its presence, and this type of information would be useful within a future longitudinal study to better understand the relationship between foot, knee, and hip symptoms.