Sleep and Pain in Osteoarthritis

Osteoarthritis is the most common type of arthritis, most often affecting the hips and knees. Apart from pain and disability, many people with osteoarthritis anecdotally report sleep disturbances. The prevalence of self-reported sleep problems in osteoarthritis varies widely ranging from 11.9% (Gore et al., 2011) to 81% (Wilcox et al., 2000), however, these figures have not been supported by studies using objective sleep measures. While the relationship between sleep and pain is bidirectional (Frohnhofen, 2018), there is increasing evidence that sleep disturbance may impact pain, more than pain impacts sleep (Finan et al., 2013). In addition, patients with hip osteoarthritis believe that if they have better subjective sleep, their pain will improve (Blagestad et al., 2016). Sleep quality before hip and knee joint replacement is also one of the most crucial factors determining both immediate and longer-term post-operative pain (Bjurström et al., 2021; Boye Larsen et al., 2021).

There are theoretical reasons to explain why poor sleep could affect pain in osteoarthritis. Fragmented sleep may activate immune and inflammatory pathways which can increase central pain processing, as well as via an effect on mood (Smith et al., 2009). Meta-analyses have shown that sleep deprivation increases pain sensitivity in healthy controls (Schrimpf et al., 2015), and both short sleep duration and insomnia are causally linked to the development of chronic widespread pain (Williams et al., 2024). In osteoarthritis, participants with both pain and insomnia have increased central sensitisation on Quantitative Sensory Testing compared to controls (Campbell et al., 2015). Emerging evidence suggests that chronotype, an individual preference for "morningness" or "eveningness" may influence sleep quality, with "morning" types reporting better sleep quality (Yazdi et al., 2014). In addition, "evening" types may be more susceptible to musculoskeletal pain (Zhang et al., 2018), although the mechanisms for this are unclear. Qualitatively, patients with osteoarthritis describe difficulty getting to sleep, due to pain and positioning problems, as well as frequent waking during the night and difficulty getting back to sleep (Whale \& Rachael, 2022). Self-reports of poor sleep in osteoarthritis are also associated with an increased level of pain catastrophizing (Campbell et al., 2015), a tendency to constantly think about pain and the negative consequences of pain (Quartana et al., 2009). A meta-analysis of plasma C-Reactive Protein (CRP) levels, a protein produced by the liver in response to the presence of inflammatory cytokines that is involved in immune activation and a biomarker for a peripheral inflammatory state, has shown a modest but significant increase in people with osteoarthritis (Jin et al., 2015). Experimental sleep deprivation is associated with higher levels of plasma CRP (Irwin et al., 2016) and cross-sectional population studies show that both shorter and longer than average sleep duration is associated with higher levels of plasma CRP (Lee et al., 2020; Zhang et al., 2023).

Potentially, as objective sleep data is time consuming and expensive to collect, most sleep studies in osteoarthritis favour subjective, self-report measures of sleep quality, such as the Pittsburgh Sleep Quality Index (PSQI) 1989) and the Insomnia Severity Index (ISI). Quantitative objective measurements, such as actigraphy or polysomnography (PSG) are much less commonly used and are typically used for only a few days. In two of the very few PSG studies of sleep in osteoarthritis, one showed more sleep-stage transitions suggestive of sleep fragmentation in people with osteoarthritis compared to controls (Yeung et al., 2018), another showed more light than deep sleep, implying worse sleep quality in people with osteoarthritis compared to controls (Leigh et al., 1988), although this study is limited by a sole recruitment of male participants. Whether this would replicate in a mixed sample remains an empirical question. In many sleep studies, it is common to find a disparity between subjective and objective measures of sleep quality, in particular, worse self-reported sleep efficiency and total sleep time, compared to objective sleep measures (Bean et al., 2021; Heitkemper et al., 2005). This phenomenon is often called 'sleep misperception' (Hughes et al., 2018).

The extent of sleep misperception varies between different clinical populations and in one study, increasing pain levels, but not age, depression or dysfunctional sleep beliefs lead to the greatest difference between subjective and objective sleep measures, suggesting that pain can cause people to negatively interpret their sleep (Hughes et al., 2018). Sleep misperception has been reported in people with several different pain conditions including rheumatoid arthritis (Hirsch et al., 1994) but not so far in osteoarthritis. Due to a paucity of empirical literature, the mechanisms underlying sleep misperception are unclear. One theory is that people who underestimate their sleep may have more frequent brief nocturnal awakenings (Harvey \& Tang, 2012), also suggested in a small objective study in osteoarthritis (Yeung et al., 2018).

Positive personality factors that may be protective of sleep or sleep misperception such as resilience or self-efficacy have not been widely studied. Resilience has been described as an important characteristic that can help individuals "bounce-back" from challenging situations allowing adaptation and personal growth and is considered important regarding outcomes related to chronic pain (Smith \& Zautra, 2008). Sleep and resilience are positively correlated in healthy individuals (Arora et al., 2022) and people with knee osteoarthritis who have high levels of psychological resilience also have higher levels of self-rated health (Hsieh et al., 2023). Three personal characteristics have been considered key characteristics of increased pain resilience - optimism, pain acceptance, and purpose in life (Smith \& Zautra, 2008). In addition, self-efficacy (Bandura, 1977) which in the context of health is a belief in one's abilities to control symptoms, as well as external factors such as emotional or social support are also likely influence recovery or "bounce-back" from setbacks. Optimists tend to report better subjective sleep (Lau et al., 2017), lower levels of clinical pain (Shanahan et al., 2021) and lower pain catastrophising scores (Hood et al., 2012). In osteoarthritis, optimism is associated with less clinical pain and less experimental pain sensitivity (Thompson et al., 2018). In addition, people with more optimistic traits tend to report better sleep quality and fewer insomnia symptoms than those with fewer optimistic traits, even if their objective sleep is impaired, demonstrating a favourable sleep misperception pattern (Hernandez et al., 2020). This suggests that optimism may be a protective factor against the negative impact of sleep misperception and subsequent pain. The broader concept of pain resilience related to sleep has yet to be investigated in clinical populations with osteoarthritis. Another factor that may contribute to sleep misperception in osteoarthritis could be pain catastrophising; a collection of negative cognitive and affective biases to pain associated with feelings of helplessness (Quartana et al., 2009). In people with osteoarthritis, sleep disturbance is associated with both pain catastrophising and the level of knee pain (Wang et al., 2023).

A systematic review looking at sleep and pain associations in osteoarthritis is currently being undertaken within the lead authors lab (Holden et al, in prep). 25 publications with a total of 13,841 participants with OA were included. The prevalence of self-reported poor sleep ranged from 25% (Gore et al., 2011) to 81.1% (Wilcox et al., 2000). Importantly, the extent to which conclusions can be drawn from the data are limited by the high heterogeneity across studies, as well as methodological limitations, including the variability in diagnostic criteria for osteoarthritis, low statistical power in objective measurement and a lack of consideration for known sleep confounders. Across the 25 included studies, 11 different subjective sleep outcome measures and 12 different pain measures were used. Objective sleep monitoring was used in only 5 studies, and of these only 2 actigraphy studies were used for 7 nights which is the generally agreed minimum number of nights needed to accurately assess sleep. 2 studies examined the effects of poor sleep on next day pain. One found an association between poor self-reported sleep, but not sleep recorded by actigraphy, and morning pain (Whibley et al., 2019). The other found no association between sleep measured by actigraphy and next day pain or vice versa (Parmelee et al., 2017). There were only 3 longitudinal studies, both of which used non-standardised methods for quantifying sleep quality. One found that more nights of restless sleep per week was associated with increased pain at baseline but not at 1 year (Lapane et al., 2021) and the other found that reports of poor sleep were also associated with baseline pain but not at 4-year follow-up (Parmelee \& Tighe, 2015). Only 2 studies considered both subjective and objective measures of sleep (Campbell et al., 2015) (Chen et al., 2015), with only one being sufficiently powered to detect an effect (Campbell et al., 2015). In this study, counterintuitively, people with osteoarthritis and insomnia had better subjective compared to objective sleep than control groups. Most cross-sectional studies reported an association between either worse sleep quality or quantity and increased pain or vice-versa, however the strengths of these associations were often weak. No studies in the systematic review considered personality factors that could have been protective of sleep.

In summary, poor sleep and pain are associated in people with osteoarthritis; however, the nature and strength of this association is unclear from the available literature. This study aims to clarify the nature and extent of sleep problems in people with clinically diagnosed hip or knee osteoarthritis using objective as well as subjective sleep measures, and to identify any sleep misperception. In addition, this study aims to clarify any association between sleep misperception and subjective pain reports and to identify factors including pain catastrophising and resilience which might mediate this association. The aim is to identify personality factors which could be amenable to a brief psychological intervention, for example mindfulness, that could influence sleep misperception and pain reporting. Measures of peripheral inflammation and central sensitisation, both theorised as mechanisms linking sleep quality and pain will be measured using plasma CRP and with Quantitative Sensory Testing.

Research Questions

The principal research question is as follows:

Is there an association between sleep misperception and pain in hip and knee osteoarthritis?

The secondary research questions are as follows:

Is negative sleep misperception associated with pain catastrophising?

Is there less sleep misperception in participants with positive personality attributes such as resilience and optimism?

Which sleep variables predict the presence/absence of sleep misperception

Is sleep misperception in osteoarthritis associated with raised levels of plasma CRP, a marker of general inflammation?

Is sleep misperception in osteoarthritis associated with central sensitisation measured with Von Frey Filaments and a thermal cutaneous stimulator?

Enrolment and Procedures Adults over 18 years old with hip and knee osteoarthritis will be recruited from NHS orthopaedic clinics at Hampshire Hospitals NHS Foundation Trust, Basingstoke. After screening for inclusion and exclusion criteria, participants will complete a consent form either on paper or via REDCap electronically. They will then complete a battery of psychometrics to quantify aspects of their personality, pain and sleep as well as their demographics. Participants will then also be asked to wear a wrist actigraphy monitor continuously for 7 days and nights to objectively measure their objective sleep quality and quantity. Additionally, each morning after sleep monitoring, participants will complete a brief sleep diary, a pain numerical rating scale and a visual analogue mood scale to assess subjective sleep quantity and quality, pain and mood over the trial period. Willing participants will also be asked to provide a blood sample of approximately 100 microlitres, via a home testing kit, collected into a capillary tube from a pinprick of blood to quantify high sensitivity C-Reactive Protein. At the conclusion of the week-long monitoring period, participants will complete several more sleep and pain questionnaires. Optional Quantitative Sensory Testing (QST) using standardised protocols will also be performed in the participants homes using Von Frey Filaments to assess pressure pain threshold and a thermal cutaneous stimulator to assess thermal pain threshold. Feedback on sleep results will be provided to each participant by telephone or videocall.