Abstract
Objective
Frailty is a multisystem syndrome and its relationship with symptomatic osteoarthritis has been reported. We aimed to identify trajectories of knee pain in a large prospective cohort and to describe the effect of frailty status at baseline on the pain trajectories over 9 years.
Methods
We included 4419 participants (mean age 61.3 years, 58% female) from the Osteoarthritis Initiative cohort. Participants were classified as “no frailty,” “pre-frailty,” or “frailty” at baseline, based on 5 characteristics (ie, unintentional weight loss, exhaustion, weak energy, slow gait speed, and low physical activity). Knee pain was evaluated annually using the Western Ontario and McMaster Universities Osteoarthritis Index pain subscale (0–20) from baseline to 9 years.
Results
Of the participants included, 38.4%, 55.4%, and 6.3% were classified as “no frailty,” “pre-frailty,” and “frailty,” respectively. Five pain trajectories were identified: “No pain” (n = 1010, 22.8%), “Mild pain” (n = 1656, 37.3%), “Moderate pain” (n = 1149, 26.0%), “Severe pain” (n = 477, 10.9%), and “Very Severe pain” (n = 127, 3.0%). Compared to participants with no frailty, those with pre-frailty and frailty were more likely to have more severe pain trajectories (pre-frailty: odds ratios [ORs] 1.5 to 2.1; frailty: ORs 1.5 to 5.0), after adjusting for potential confounders. Further analyses indicated that the associations between frailty and pain were mainly driven by exhaustion, slow gait speed, and weak energy.
Conclusions
Approximately two-thirds of middle-aged and older adults were frail or pre-frail. The role of frailty in predicting pain trajectories suggests that frailty may be an important treatment target for knee pain.
Keywords: frailty, knee pain, osteoarthritis, trajectory
Introduction
Osteoarthritis (OA) is the most common joint disease and source of pain in older adults.1 As the core clinical feature of OA, pain is the primary reason for patients seeking medical attention.2 Despite this, pain control is inadequate in over half of the patients.3 Given that OA may respond better to early prevention and treatment, it is important to identify patients at higher risks of progressing to severe pain.4
Frailty is a multisystem syndrome characterized by vulnerability to stressors and decreased physiological reserves, yielding a limited capacity to maintain homeostasis.5 Frailty can be assessed clinically using the “frailty index score” based on a broad spectrum of health deficits of an individual,6 or using the “frailty phenotypes” based on 5 factors (unintentional weight loss, exhaustion, weakness, slow gait speed, and low physical activity).7 Frailty predicts increased risks of falls, disability, hospitalization, and mortality.7 Epidemiologic studies have indicated that symptomatic OA and other chronic pain were associated with an increased risk of frailty.8,9 The association is due to that pain and frailty may share common mechanisms, including inflammation, neuroinflammation, and endocrine dysregulation.10 While current studies have focused on the effect of pain on frailty, it has been shown that frailty may be a risk factor for the development of OA by increasing muscle atrophy, inflammation, and fall-related injury.11 This suggests that frailty could contribute to the progression of pain and be a treatment target. Moreover, while pain in OA patients frequently gets worse over time at the group level, recent work has found different courses for the progression of pain (eg, worsening, improving, or stable) at the individual level.12–14 Identifying pain trajectories and factors related to undesirable trajectories may be critical for the prevention and management of OA. Thus, this study aimed to identify pain trajectories in a large prospective cohort and to describe the role of frailty at baseline in predicting pain trajectories over 9 years.
Methods
Study design
This study was reported according to the STROBE statement.15 The Osteoarthritis Initiative (OAI) is a publicly available, multicenter, prospective cohort study conducted in 4 clinical centers in the United States.16 The OAI comprised 4796 participants aged 45–79 years with or at increased risk of knee OA (https://nda.nih.gov/oai). Participants were excluded from the OAI if they had inflammatory arthritis, bilateral end-stage knee OA, bilateral knee replacement surgeries, or had contraindications to 3-T magnetic resonance imaging (MRI). Ethics approval was obtained from the institutional review board at four clinical centres where participants were recruited. All participants provided written informed consent. In this study, we included 4419 participants who had sufficient data to define frailty status at baseline.
Outcome assessment: knee pain
Knee pain in both knees was assessed at baseline (Feb 2004—May 2006) and the annual follow-up visits over 9 years (10 assessments in total) using the 5-item Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain subscale (total score: 0 [no pain] to 20 [extreme pain]), which measures pain during 5 activities (ie, flat surface walking, stair climbing, at night, sitting or lying, standing).17 The most painful knee was used to identify pain trajectories in this study.
Exposure assessment: frailty
Frailty status was assessed according to the frailty phenotype criteria, which include 5 components (total score 0–5): unintentional weight loss, exhaustion, weak energy, slow gait speed, and low physical activity.7 The ascertainment of positive components of the frailty phenotype criteria was consistent with a previous study that used data from the OAI cohort.8 Specifically, weight loss was considered positive if there was an unintentional weight loss of ≥ 5% from baseline to the 12-month follow-up. Exhaustion was assessed using 2 items from the Center for Epidemiologic Studies-Depression (CES-D) scale: “How often have you felt that everything required considerable effort during the past week?” and “How often could you not get going during the past week?”18 The answers to these questions were: 0—rarely or none of the time (less than one day), 1—some of the time (one to two days), 2—much of the time (three to four days), and 3—most of the time (five to seven days). Exhaustion was considered positive if participants answered 2 or 3 to either of the 2 questions. Because handgrip strength was not measured in the OAI study, the assessment of weak energy was substituted using 1 item from the 12-item short-form health survey: “How often have you had a lot of energy in the past four weeks.”19 The answers were: 1—all of the time, 2—most of the time, 3—some of the time, 4—a little of the time, and 5—none of the time. Weak energy was considered positive if participants answered 4 or 5 to that question. Gait speed (m/s) was assessed using a standard 20-meter walk test,20 and slow gait speed was defined as gait speed slower than 1 m/s. Physical activity was assessed using the Physical Activity Scale for the Elderly (PASE),21 and participants with the lowest 20% of the PASE score, adjusted for sex, were considered to have low physical activity. Pre-frailty and frailty were considered positive if the total score of the frailty phenotype criteria was 1–2 and ≥3, respectively, and a score of 0 indicates no frailty.7 To minimize the impact of missing data on frailty components, we ascertained the frailty status if the missingness did not influence the classification of frailty status. For example, participants with a total frailty score of 1 and had only 1 missing component would be considered pre-frail, because the missing component would not change the frailty status (ie, total frailty scores of both 1 and 2 are classified as pre-frailty).
Assessments of covariates
Covariates were selected based on previous literature and a directed acyclic graph (Supplementary Figure 1) for their associations with both frailty and knee pain trajectory.14,22–25 They included the following variables: age (year), sex, self-reported race (White, Black or African Americans, Other), body mass index (BMI, kg/m2), education level (less than high school, high school, some college or associate degree or technical school after completing high school, college graduate, some graduate school, graduate degree), income (<10k, 10–25k, 25–50k, 50–100k, >100k), depression (CES-D ≥ 16),26 Charlson comorbidity score (0, 1–2, ≥3),27 pain-relief treatments (acetaminophen, non-steroidal anti-inflammatory drugs, analgesics, intra-articular steroids, and non-medication therapies such as acupuncture and massage), and radiographic OA (ROA). ROA was defined as a Kellgren-Lawrence grade (KLG) ≥2 in either knee using a fixed flexion knee X-ray.28 Two projects in the OAI (Projects 15 and 37/42) were conducted to measure KLG, with some duplicated readings. The “worst” KLG from the 2 projects was used,29 as recommended by the OAI handbook.
Statistical analysis
Baseline characteristics of participants were described using mean (standard deviation), n (%), or median (interquartile range), split by frailty status (ie, no frailty, pre-frailty, and frailty) and as a whole.
Group-based trajectory modelling (GBTM) was used to identify knee pain trajectories that shared similar courses of pain evolution over the 9 years of follow-up.30 GBTM is widely used to identify homogeneous clusters of developmental trajectories within the population.31 Participants were assigned to a certain trajectory that they had the highest probability of belonging; such that different trajectories reflect distinctive aetiologies. For each of the two to six trajectories, we applied censored normal models starting with a linear form for all trajectory groups, and then we increased the complexity of the shape of trajectories (ie, from linear to quadratic and cubic) until there was no further improvement in model fit. The selection of the number of groups and shape of trajectories was based on 2 prespecified criteria: (1) model fit: lower absolute values of the Bayesian Information Criterion (BIC) indicated a better model fit, where an increase in BIC above 10 points was considered a significant improvement in model fit;32 and (2) model distinctiveness, interpretability, and clinical relevance. Based on previous literature, more parsimonious models that included fewer trajectory groups and simpler shapes of the trajectory (eg, linear over quadratic) were chosen.32 The minimum average posterior probability (entropy) of assigning participants to each trajectory was set at 0.7, and the minimum proportion of participants classified in each group with the minimum posterior probability of 0.7 was set at 70%, according to the recommended criteria.31 The odds of correct classification (OCC) was calculated to assess the accuracy of trajectory models, which is the ratio of the odds of correct classification on the basis of the maximum probability classification rule to the odds of correct classification based on random assignment.33 An OCC of >5 was required for each trajectory group.31 Individual trajectories for each participant were plotted and fitted using Locally Weighted Scatterplot Smoothing (LOWESS), by trajectory group.
Multinomial logistic regression models were used to evaluate the association of baseline frailty status and frailty components with distinctive pain trajectories using the “No pain” trajectory as the referent group. Apart from univariable models, we built multivariable models to adjust for baseline covariates including age, sex, race, BMI, education level, income, comorbidity, and ROA. Odds ratios (OR) with 95% confidence intervals (CI) were calculated to describe these associations.
Several sensitivity analyses were performed. First, to address missing data on covariates at baseline (0.02% to 6.4% missing), multiple imputations with chained equations were performed with 20 imputations using complete variables (age, sex, frailty status, and pain trajectories) and nonmissing values of covariates at baseline, assuming missing at random. Second, we rerun the analyses using data of participants who had all 10 measurements of knee pain. Third, in view of the stability of pain trajectories identified in this study, we explored the association between frailty status and pain trajectories by further adjusting for knee pain at baseline. All data analyses were performed using Stata version 16.1 (StataCorp, TX). A two-sided P values less than 0.05 was considered statistically significant.
Results
Participants
Of 4419 participants included (mean age 61.3 years, 58% female), 1695 (38.4%) were not frail, 2446 (55.4%) were pre-frail, and 278 (6.3%) were frail. Table 1 summarises the baseline characteristics of study participants. Compared to participants without frailty at baseline, those with pre-frailty and frailty were older, more likely to be females and Black or African Americans, had higher BMI, more severe knee pain, ROA, and more comorbidities, and took more pain-relief medications. Many participants with pre-frailty and frailty were positive for exhaustion (68% and 94%, respectively), followed by weak energy (29% and 77%, respectively).
Table 1.
Baseline characteristics of study participants
| Overall | No frailty | Pre-frailty | Frailty | |
|---|---|---|---|---|
| (n = 4419) | (n = 1695) | (n = 2446) | (n = 278) | |
| Age, years, mean (SD) | 61.3 (9.2) | 60.7 (8.9) | 61.5 (9.2) | 62.9 (9.9) |
| Female | 2570 (58) | 947 (56) | 1279 (58) | 180 (67) |
| Race | ||||
| White participants | 3578 (81) | 1476 (87) | 1927 (79) | 164 (59) |
| Black participants | 741 (17) | 182 (11) | 454 (19) | 105 (38) |
| Other participants | 106 (2) | 36 (2) | 61 (2) | 9 (3) |
| Body mass index, kg/m2, mean (SD) | 28.6 (4.8) | 27.7 (4.4) | 28.9 (4.9) | 31.4 (5.4) |
| WOMAC pain (range 0–20), median (IQR) | 2 (0–5) | 1 (0–4) | 3 (1–6) | 6 (3–10) |
| Radiographic osteoarthritis | 2414 (56) | 894 (53) | 1350 (58) | 170 (69) |
| Education level | ||||
| Less than high school | 140 (3) | 25 (1) | 84 (3) | 31 (11) |
| High school | 546 (12) | 168 (10) | 309 (13) | 69 (25) |
| Some college | 1044 (24) | 339 (20) | 627 (26) | 78 (28) |
| College graduate | 946 (21) | 413 (24) | 495 (20) | 38 (13) |
| Some graduate school | 376 (9) | 148 (9) | 215 (9) | 13 (5) |
| Graduate degree | 1366 (31) | 602 (36) | 715 (29) | 49 (18) |
| Income, dollar | ||||
| <10k | 138 (3) | 25 (2) | 72 (3) | 42 (17) |
| 10–25k | 408 (10) | 100 (6) | 256 (11) | 52 (21) |
| 25–50k | 1047 (25) | 327 (20) | 643 (28) | 77 (31) |
| 50–100k | 1520 (37) | 639 (40) | 827 (36) | 54 (22) |
| >100k | 1021 (25) | 511 (32) | 489 (21) | 21 (9) |
| Depression | 434 (10) | 10 (0.6) | 308 (13) | 116 (42) |
| Comorbidity | ||||
| No | 3334 (76) | 1380 (82) | 1804 (74) | 150 (56) |
| 1–2 | 925 (21) | 280 (16) | 555 (23) | 90 (33) |
| 3 or more | 135 (3) | 30 (2) | 75 (3) | 30 (11) |
| Pain-relief treatments | ||||
| Acetaminophen | 454 (10) | 134 (8) | 255 (10) | 65 (23) |
| NSAIDs | 1345 (30) | 420 (25) | 792 (32) | 133 (48) |
| Analgesics | 115 (3) | 17 (1) | 72 (3) | 26 (9) |
| Intra-articular steroids | 81 (2) | 21 (1) | 47 (2) | 13 (5) |
| Non-medication therapies | 450 (10) | 152 (10) | 272 (11) | 27 (10) |
| Frailty components | ||||
| Weight loss ≥ 5% | 383 (9) | 0 | 305 (13) | 78 (34) |
| Exhaustion (2 items of the CES-D) | 1912 (43) | 0 | 1652 (68) | 260 (94) |
| Weak energy (1 item of the SF-12) | 912 (21) | 0 | 697 (29) | 215 (77) |
| Gait speed < 1 m/s | 290 (7) | 0 | 147 (6) | 143 (52) |
| Low activity (lowest 20% of the PASE) | 492 (11) | 0 | 293 (12) | 199 (72) |
CES-D = Center for Epidemiological Studies-Depression; IQR = interquartile range; NSAIDs = non-steroidal anti-inflammatory drugs; PASE = Physical Activity Scale for the Elderly; SD = standard deviation; SF-12 = 12-Item Short-Form Survey; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index.
Results are shown as n (%) unless stated otherwise (ie, mean [SD], median [IQR]).
Based on BIC, average posterior probability, the proportion of participants with posterior probability >0.7, and OCC (Supplementary Table 1), 5 pain trajectories with distinct pain severity and similar shape were identified (Figure 1). Based on the severity of the identified pain trajectories, they were named “No pain” (n = 1010, 22.8%), “Mild pain” (n = 1656, 37.3%), “Moderate pain” (n = 1149, 26.0%), “Severe pain” (n = 477, 10.9%), and “Very Severe pain” (n = 127, 3.0%). Figure 2 shows the individual trajectories of participants in each of the 5 trajectories.
Figure 1.
Trajectory of knee pain severity measured by the WOMAC pain subscale (0–20, higher score indicates more severe pain) over 9 years.
Figure 2.
Individual trajectories of knee pain severity measured by the WOMAC pain subscale (0–20, higher score indicates more severe pain) over 9 years. Red indicates fitted lines.
Association of frailty status and pain trajectories
In both univariable and multivariable analyses, participants with pre-frailty and frailty, compared to those with no frailty, were more likely to have more severe pain trajectories (pre-frailty: OR 1.5–2.1; frailty: OR 1.5–5.0; Table 2). Moreover, there was a statistically significant dose-response relationship between the severity of frailty and the odds of more severe pain trajectories (p for trend < 0.001). In the analyses of the associations between frailty components and pain trajectories, exhaustion (OR 1.7–3.2), slow gait speed (OR 1.1–4.0), and weak energy (OR 0.9–1.6) were significantly associated with more severe pain trajectories, and low physical activity was significantly associated with a reduced risk of “Moderate pain” trajectory (OR 0.7, 95% CI 0.6–0.9; Table 3).
Table 2.
Association between frailty status at baseline and knee pain trajectories
| Odds Ratio (95% Confidence Interval)a |
||||
|---|---|---|---|---|
| Mild pain | Moderate pain | Severe pain | Very Severe pain | |
| Univariable model (n = 4419) | ||||
| No frailty | Reference | Reference | Reference | Reference |
| Pre-frailty | 1.6 (1.3 to 1.8) | 2.1 (1.7 to 2.5) | 3.3 (2.5 to 4.2) | 3.7 (2.2 to 6.4) |
| Frailty | 1.9 (1.2 to 3.1) | 4.3 (2.7 to 6.9) | 15.9 (9.6 to 26.3) | 51.6 (26.3 to 101.3) |
| P for trend | <.001 | <.001 | <.001 | <.001 |
| Multivariable model (n = 3981) | ||||
| No frailty | Reference | Reference | Reference | Reference |
| Pre-frailty | 1.5 (1.2 to 1.8) | 1.6 (1.3 to 2.0) | 2.1 (1.6 to 2.9) | 1.8 (0.9 to 3.4) |
| Frailty | 1.5 (0.8 to 2.6) | 2.1 (1.1 to 3.7) | 4.3 (2.2 to 8.4) | 5.0 (1.9 to 12.7) |
| P for trend | <.001 | <.001 | <.001 | <.001 |
Multivariable model adjusted for age, sex, race, body mass index, education level, income, depression, comorbidity, pain-relief treatments (acetaminophen, non-steroidal anti-inflammatory drugs, analgesics, intra-articular steroids, and non-medication therapies such as acupuncture and massage), and radiographic knee osteoarthritis.
No pain was set as the referent trajectory.
Bold denotes statistically significant results.
Table 3.
Associations between frailty components and knee pain trajectories
| Odds Ratio (95% Confidence Interval)a |
||||
|---|---|---|---|---|
| Mild pain | Moderate pain | Severe pain | Very Severe pain | |
| Univariable model (n = 4176) | ||||
| Unintentional weight loss | 1.3 (0.9 to 1.7) | 1.2 (0.9 to 1.7) | 1.7 (1.1 to 2.5) | 1.9 (1.0 to 3.6) |
| Exhaustion | 1.6 (1.4 to 2.0) | 2.3 (1.9 to 2.8) | 3.2 (2.5 to 4.1) | 2.8 (1.8 to 4.4) |
| Weak energy | 1.2 (0.8 to 1.7) | 2.2 (1.5 to 3.1) | 3.2 (2.2 to 4.7) | 4.7 (2.8 to 8.1) |
| Slow gait speed | 1.3 (0.8 to 2.1) | 2.6 (1.6 to 4.2) | 4.4 (2.7 to 7.4) | 13.1 (7.2 to 23.8) |
| Low physical activity | 1.1 (0.9 to 1.4) | 0.9 (0.7 to 1.1) | 1.1 (0.8 to 1.5) | 1.5 (0.9 to 2.4) |
| Multivariable model (n = 3877) | ||||
| Unintentional weight loss | 1.2 (0.9 to 1.7) | 0.9 (0.6 to 1.4) | 1.3 (0.8 to 2.0) | 1.2 (0.6 to 2.5) |
| Exhaustion | 1.7 (1.4 to 2.1) | 2.3 (1.8 to 2.8) | 3.2 (2.4 to 4.3) | 2.2 (1.3 to 3.9) |
| Weak energy | 0.9 (0.6 to 1.3) | 1.4 (0.9 to 2.0) | 1.6 (1.0 to 2.6) | 1.6 (0.8 to 3.3) |
| Slow gait speed | 1.1 (0.6 to 2.0) | 1.8 (1.0 to 3.2) | 2.1 (1.1 to 3.9) | 4.0 (1.9 to 8.8) |
| Low physical activity | 1.0 (0.8 to 1.2) | 0.7 (0.6 to 0.9) | 0.8 (0.6 to 1.1) | 0.9 (0.5 to 1.6) |
Multivariable model adjusted for age, sex, race, body mass index, education level, income, depression, comorbidity, pain-relief treatments (acetaminophen, non-steroidal anti-inflammatory drugs, analgesics, intra-articular steroids, and non-medication therapies such as acupuncture and massage), and radiographic knee osteoarthritis.
No pain was set as the referent trajectory.
All 5 frailty components were simultaneously included in the same models (univariable/multivariable models). Bold denotes statistically significant results.
Sensitivity analysis
Multiple imputations for missing data of covariates did not materially change the association between frailty status and knee pain trajectories (Supplementary Table 2). Moreover, similar pain trajectories and associations with frailty status were also found when restricting to participants with all 10 measurements of knee pain (n = 3010, Supplementary Figure 2 and Supplementary Table 3). Further adjustment for baseline knee pain weakened the association between frailty and pain trajectories but their association remained statistically significant (Supplementary Table 4).
Discussion
This is the first study to evaluate the role of frailty in the progression of knee pain. Using data from the OAI, we identified 5 distinct but stable pain trajectories and found that pre-frailty and frailty were common in this population of middle-aged and older adults with or at increased risk of knee OA. Importantly, both pre-frailty and frailty were associated with more severe pain trajectories over the 9 years of follow-up, for which the main drivers may be exhaustion and slow gait speed. These findings suggest that frailty plays an important role in the progression of knee pain, and that further studies are needed to evaluate whether treatments targeting frailty components, especially exhaustion and slow gait speed, is beneficial for the management of pain.
The 5 pain trajectories identified in this study were consistent with previous studies of community-dwelling older adults that identified 3–5 pain trajectories.12–14,34 Overall, the shapes of pain trajectories identified in previous studies and the present study were stable, suggesting that knee pain tends to be non-progressive and maintained over time in middle-aged and older adults with or at increased risk of knee OA. Moreover, the prevalence of frailty and pre-frailty were 6.3% and 55.4%, respectively, and the characteristics of frailty people in this OAI population were similar to other populations.7,35 For example, frailty and pre-frailty were more commonly seen in older adults, women, Black or African Americans, and those with more comorbidities. This provides an indication for screening frail people in clinical practice.
Our findings showed that, compared to participants without frailty, those with either frailty or pre-frailty were more likely to have more severe pain trajectories, even after controlling for knee pain at baseline. While frail people have a 5- to 6-fold risk of “Severe” and “Very Severe” pain trajectories compared to people with no frailty, people at the intermediate stage of frailty (ie, pre-frailty) also had a significantly increased risk of severe pain trajectories. Previous studies have shown that pain was a risk factor for frailty,9 and this study adds new knowledge that frailty, in turn, predicts pain trajectories in middle-aged and older adults. Therefore, screening for both frail and pre-frail people among middle-aged and older adults may be critical to identify those at high risks of undesirable pain trajectories. However, the cost-effectiveness should be evaluated before implementing such screening in a wide range. The association between frailty and knee pain trajectory provides new insights into the management of osteoarthritic pain, and the effects of existing treatments on knee pain could also be influenced by frailty status, which should be considered in clinical studies and clinical practice. At the population level, reversing frailty status may have additional public health significance in the large number of older adults who are suffering from knee pain. In addition, because of the high prevalence of pre-frailty (ie, almost 1.5 times as many as people with no evidence of frailty), targeting adults with pre-frailty can be equally or even more important for the management of pain. Previous studies have shown that exercise and nutritional interventions can reverse frailty and corresponding outcomes,36–39 and guidelines for the treatment of frailty have been built.40 Further studies are needed to determine the effect of frailty-targeted treatments on pain relief as well as the moderating effect of frailty on treatments targeting knee pain.
There are several potential mechanisms underlying the association between frailty status and worse pain trajectories. Frailty is a physiologic syndrome that preludes to disability.7 The progressive disability can reduce daily physical activity, thereby leading to worse pain trajectories.41 A previous study has shown that pro-inflammatory cytokines are elevated in frail individuals, which may accumulate in the synovial joints and cause inflammation-related joint pain.11,42 Moreover, frail individuals have worse muscle atrophy and less stable joints, which can result in increased susceptibility to biomechanical damage.43 The increased risk of falls in frail individuals may also increase the risk of post-traumatic pain.7
Among the 5 frailty components, we found that exhaustion, slow gait speed, and possibly weak energy showed strong and statistically significant associations with more severe pain trajectories. This agrees with previous findings that chronic fatigue syndrome, a condition that causes exhaustion and weak energy, is closely related to muscle and joint pain throughout the body.44 Moreover, a recent study also indicated that greater muscle strength and quality were protective of the progression of knee pain, and that improving muscle function and composition may be important for altering unfavourable knee pain trajectories.45 The association between slow gait speed and knee pain has long been observed, and it is more likely that pain leads to declines in gait speed.46,47 However, slow gait speed also reflects a worse health condition, including poor balance and muscle function, which could be improved through exercise and physiotherapy.48,49 We found that low physical activity showed a trend to be associated with reduced risks of worse pain trajectories, and this was statistically significant for the “Moderate pain” trajectory. This could be a chance finding since physical activity has been shown to be beneficial for knee pain.41 Despite this, another study also indicated that increases in daily physical activity were associated with increased pain in knee OA patients.50
The strengths of this study include the large sample size and ten repeated measurements of knee pain over a long follow-up. There are several limitations in this study. First, we cannot determine whether weight loss, as a component of frailty, was unintentional or not. This may have overestimated the prevalence of frailty and pre-frailty. Moreover, weak energy was assessed based on 1 item from the 12-item short-form health survey rather than grip strength, as defined by Fried and colleagues.7 Thus, misclassification may have occurred, and further studies are needed to verify our findings. Second, while the pain trajectories identified in this study were similar to previous studies12–14,34 and the model fit was good and assignment accuracy high, these trajectories were relatively stable and did not reflect participants with a worsening, improvement, or fluctuation of pain. This is due to that the GBTM assumes that individual deviations within each of the identified trajectories were due to random error. However, a previous study using latent class growth mixture models, which can capture within-group variations, had similar findings to our study, suggesting that our findings are robust.
In conclusion, almost two-thirds of middle-aged and older adults were frail or pre-frail in this community-dwelling cohort. The role of frailty in predicting pain trajectories suggests that frailty may be an important treatment target for knee pain.
Supplementary Material
Acknowledgments
We thank the staff and volunteers of the OAI study. The OAI is a publicly available, multicentre, prospective cohort study conducted in four centres in the US.
Author contributions: G.C, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study design: G.C., X.P., and F.P. Data clean: Y.W., Y.Z., and X.L. Analysis and interpretation of data: G.C., Y.Z., Y.W., X.L., S.X., Z.S., F.P., and X.P. Manuscript preparation, revision, and approval: G.C., Y.Z., Y.W., X.L., S.X., Z.S., F.P., and X.P.
Statement of ethics: Ethics approval was obtained from the institutional review board at four clinical centres where participants were recruited. All participants provided written informed consent.
Supplementary material
Supplementary material is available at Pain Medicine online.
Contributor Information
Guoqi Cai, Department of Epidemiology and Biostatistics, School of Public Health, Anhui Medical University, Hefei, Anhui, China; Menzies Institute for Medical Research, University of Tasmania, Hobart, TAS, Australia.
Youyou Zhang, Department of Epidemiology and Biostatistics, School of Public Health, Anhui Medical University, Hefei, Anhui, China.
Yining Wang, Department of Epidemiology and Biostatistics, School of Public Health, Anhui Medical University, Hefei, Anhui, China.
Xiaoxi Li, Department of Epidemiology and Biostatistics, School of Public Health, Anhui Medical University, Hefei, Anhui, China.
Shengqian Xu, Department of Rheumatism and Immunity, The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, China.
Zongwen Shuai, Department of Rheumatism and Immunity, The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, China.
Faming Pan, Department of Epidemiology and Biostatistics, School of Public Health, Anhui Medical University, Hefei, Anhui, China.
Xiaoqing Peng, Department of Obstetrics and Gynaecology, The First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, China; School of Pharmacology, Anhui Medical University, Hefei, Anhui, China; NHC Key Laboratory of Study on Abnormal Gametes and Reproductive Tract (Anhui Medical University), Hefei, Anhui, China.
Funding
The preparation of this work was supported by the National Natural Science Foundation of China (82103933, 81773514, 82073655). The study and image acquisition was funded by the OAI, a public-private partnership composed of 5 contracts (N01-AR-2–2258; N01-AR-2–2259; N01-AR-2–2260; N01-AR-2–2261; N01-AR-2–2262) funded by the National Institutes of Health, a branch of the Department of Health and Human Services, and conducted by the OAI Study Investigators. Private funding partners include Merck Research Laboratories; Novartis Pharmaceuticals Corporation, GlaxoSmithKline and Pfizer, Inc. Private sector funding for the OAI is managed by the Foundation for the National Institutes of Health. This manuscript was prepared using an OAI public use data set and does not necessarily reflect the opinions or views of the OAI investigators, the National Institutes of Health (NIH), or the private funding partners. None of the study sponsors had any role in data collection, storage or analysis, in manuscript writing, or the decision to publish this manuscript.
Conflicts of interest: The authors declare that they have no conflicts of interest.
Data availability
The data that support the findings of this study are openly available at https://nda.nih.gov/oai.
References
- 1. Shane Anderson A, Loeser RF.. Why is osteoarthritis an age-related disease? Best Pract Res Clin Rheumatol. 2010;24(1):15–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Hadler NM. Knee pain is the malady–not osteoarthritis. Ann Intern Med. 1992;116(7):598–599. [DOI] [PubMed] [Google Scholar]
- 3. Conaghan PG, Peloso PM, Everett SV, et al. Inadequate pain relief and large functional loss among patients with knee osteoarthritis: Evidence from a prospective multinational longitudinal study of osteoarthritis real-world therapies. Rheumatology (Oxford). 2015;54(2):270–277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Roos EM, Arden NK.. Strategies for the prevention of knee osteoarthritis. Nat Rev Rheumatol. 2016;12(2):92–101. [DOI] [PubMed] [Google Scholar]
- 5. Fedarko NS. The biology of aging and frailty. Clin Geriatr Med. 2011;27(1):27–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Rockwood K, Mitnitski A.. Frailty in relation to the accumulation of deficits. J Gerontol A Biol Sci Med Sci. 2007;62(7):722–727. [DOI] [PubMed] [Google Scholar]
- 7. Fried LP, Tangen CM, Walston J, et al. ; Cardiovascular Health Study Collaborative Research Group. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146–56. [DOI] [PubMed] [Google Scholar]
- 8. Bindawas SM, Vennu V, Stubbs B.. Longitudinal relationship between knee pain status and incident frailty: data from the osteoarthritis initiative. Pain Med. 2018;19(11):2146–2153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Saraiva MD, Suzuki GS, Lin SM, de Andrade DC, Jacob-Filho W, Suemoto CK.. Persistent pain is a risk factor for frailty: a systematic review and meta-analysis from prospective longitudinal studies. Age Ageing. 2018;47(6):785–793. [DOI] [PubMed] [Google Scholar]
- 10. D'Agnelli S, Amodeo G, Franchi S, et al. Frailty and pain, human studies and animal models. Ageing Res Rev. 2022;73:101515. [DOI] [PubMed] [Google Scholar]
- 11. O'Brien MS, McDougall JJ.. Age and frailty as risk factors for the development of osteoarthritis. Mech Ageing Dev. 2019;180:21–28. [DOI] [PubMed] [Google Scholar]
- 12. Collins JE, Katz JN, Dervan EE, Losina E.. Trajectories and risk profiles of pain in persons with radiographic, symptomatic knee osteoarthritis: data from the osteoarthritis initiative. Osteoarthritis Cartilage. 2014;22(5):622–630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Nicholls E, Thomas E, van der Windt DA, Croft PR, Peat G.. Pain trajectory groups in persons with, or at high risk of, knee osteoarthritis: findings from the knee clinical assessment study and the osteoarthritis initiative. Osteoarthritis Cartilage. 2014;22(12):2041–2050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Pan F, Tian J, Aitken D, Cicuttini F, Jones G.. Predictors of pain severity trajectory in older adults: a 10.7-year follow-up study. Osteoarthritis Cartilage. 2018;26(12):1619–1626. [DOI] [PubMed] [Google Scholar]
- 15. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med. 2007;147(8):573–577. [DOI] [PubMed] [Google Scholar]
- 16. Peterfy CG, Schneider E, Nevitt M.. The osteoarthritis initiative: report on the design rationale for the magnetic resonance imaging protocol for the knee. Osteoarthritis Cartilage. 2008;16(12):1433–1441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW.. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol. 1988;15(12):1833–1840. [PubMed] [Google Scholar]
- 18. Radloff L. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Measur. 1977;1(3):385–401. [Google Scholar]
- 19. Ware J Jr., Kosinski M, Keller SD.. A 12-item short-form health survey: construction of scales and preliminary tests of reliability and validity. Med Care. 1996; 34(3):220–233. [DOI] [PubMed] [Google Scholar]
- 20. Motyl JM, Driban JB, McAdams E, Price LL, McAlindon TE.. Test-retest reliability and sensitivity of the 20-meter walk test among patients with knee osteoarthritis. BMC Musculoskelet Disord. 2013;14:166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Washburn RA, Smith KW, Jette AM, Janney CA.. The Physical Activity Scale for the Elderly (PASE): development and evaluation. J Clin Epidemiol. 1993;46(2):153–162. [DOI] [PubMed] [Google Scholar]
- 22. Feng Z, Lugtenberg M, Franse C, et al. Risk factors and protective factors associated with incident or increase of frailty among community-dwelling older adults: a systematic review of longitudinal studies. PLoS One. 2017;12(6):e0178383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Yang F, Chen QW.. Evaluation of frailty and influencing factors in old people in hospital institution: evidence for a phenotype of frailty. Medicine (Baltimore). 2018;97(3):e9634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Legha A, Burke DL, Foster NE, et al. Do comorbidities predict pain and function in knee osteoarthritis following an exercise intervention, and do they moderate the effect of exercise? Analyses of data from three randomized controlled trials. Musculoskelet Care. 2020;18(1):3–11. [DOI] [PubMed] [Google Scholar]
- 25. Lee JY, Han K, Park YG, Park SH.. Effects of education, income, and occupation on prevalence and symptoms of knee osteoarthritis. Sci Rep. 2021;11(1):13983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Weissman MM, Sholomskas D, Pottenger M, Prusoff BA, Locke BZ.. Assessing depressive symptoms in five psychiatric populations: a validation study. Am J Epidemiol. 1977;106(3):203–214. [DOI] [PubMed] [Google Scholar]
- 27. Katz JN, Chang LC, Sangha O, Fossel AH, Bates DW.. Can comorbidity be measured by questionnaire rather than medical record review? Med Care. 1996;34(1):73–84. [DOI] [PubMed] [Google Scholar]
- 28. Kellgren JH, Lawrence JS.. Radiological assessment of osteo-arthrosis. Ann Rheum Dis. 1957;16(4):494–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Cai G, Li X, Zhang Y, et al. Knee symptom but not radiographic knee osteoarthritis increases the risk of falls and fractures: results from the Osteoarthritis Initiative. Osteoarthritis Cartilage. 2022;30(3):436–442. [DOI] [PubMed] [Google Scholar]
- 30. Jones BL, Nagin DS.. A note on a stata plugin for estimating group-based trajectory models. Sociol Methods Res. 2013;42(4):608–613. [Google Scholar]
- 31. Nagin DS. Group-Based Modeling of Development. Harvard University Press; 2005. [Google Scholar]
- 32. Frankfurt S, Frazier P, Syed M, Jung KR.. Using group-based trajectory and growth mixture modeling to identify classes of change trajectories. Counsel Psychol. 2016;44(5):622–660. [Google Scholar]
- 33. Zheng Y, Song M, Manson JE, Giovannucci EL, Hu FB.. Group-based trajectory of body shape from ages 5 to 55 years and cardiometabolic disease Risk in 2 US cohorts. Am J Epidemiol. 2017;186(11):1246–1255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Dai Z, Lu N, Niu J, Felson DT, Zhang Y.. Dietary Fiber Intake in Relation to Knee Pain Trajectory. Arthritis Care Res (Hoboken). 2017;69(9):1331–1339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Castell MV, van der Pas S, Otero A, et al. Osteoarthritis and frailty in elderly individuals across six European countries: results from the European Project on OSteoArthritis (EPOSA). BMC Musculoskelet Disord. 2015;16(1):359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Theou O, Stathokostas L, Roland KP, et al. The effectiveness of exercise interventions for the management of frailty: a systematic review. J Aging Res. 2011;2011:569194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Clegg AP, Barber SE, Young JB, Forster A, Iliffe SJ.. Do home-based exercise interventions improve outcomes for frail older people? Findings from a systematic review. Rev Clin Gerontol. 2012;22(1):68–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Kelaiditi E, van Kan GA, Cesari M.. Frailty: role of nutrition and exercise. Curr Opin Clin Nutr Metab Care. 2014;17(1):32–39. [DOI] [PubMed] [Google Scholar]
- 39. Tarazona-Santabalbina FJ, Gomez-Cabrera MC, Perez-Ros P, et al. A multicomponent exercise intervention that reverses frailty and improves cognition, emotion, and social networking in the community-dwelling frail elderly: a randomized clinical trial. J Am Med Dir Assoc. 2016;17(5):426–433. [DOI] [PubMed] [Google Scholar]
- 40. Fairhall N, Langron C, Sherrington C, et al. Treating frailty–a practical guide. BMC Med. 2011;9:83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. McAlindon TE, Bannuru RR, Sullivan MC, et al. OARSI guidelines for the non-surgical management of knee osteoarthritis. Osteoarthritis Cartilage. 2014;22(3):363–388. [DOI] [PubMed] [Google Scholar]
- 42. Giovannini S, Onder G, Liperoti R, et al. Interleukin-6, C-reactive protein, and tumor necrosis factor-alpha as predictors of mortality in frail, community-living elderly individuals. J Am Geriatr Soc. 2011;59(9):1679–1685. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Misra D, Felson DT, Silliman RA, et al. Knee osteoarthritis and frailty: findings from the multicenter osteoarthritis study and osteoarthritis initiative. J Gerontol A Biol Sci Med Sci. 2015;70(3):339–344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Snijders GF, van den Ende CH, Fransen J, et al. ; Nijmegen OsteoArthritis Collaboration Study Group Fatigue in knee and hip osteoarthritis: the role of pain and physical function. Rheumatology (Oxford). 2011;50(10):1894–1900. [DOI] [PubMed] [Google Scholar]
- 45. Pan F, Tian J, Scott D, Cicuttini F, Jones G.. Muscle function, quality, and relative mass are associated with knee pain trajectory over 10.7 years. Pain. 2022;163(3):518–525. [DOI] [PubMed] [Google Scholar]
- 46. White DK, Niu J, Zhang Y.. Is symptomatic knee osteoarthritis a risk factor for a trajectory of fast decline in gait speed? Results from a longitudinal cohort study. Arthritis Care Res (Hoboken). 2013;65(2):187–194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Kitayuguchi J, Kamada M, Hamano T, et al. Association between knee pain and gait speed decline in rural Japanese community-dwelling older adults: 1-year prospective cohort study. Geriatr Gerontol Int. 2016;16(1):55–64. [DOI] [PubMed] [Google Scholar]
- 48. Tani K, Kola I, Dhamaj F, Shpata V, Zallari K.. Physiotherapy effects in gait speed in patients with knee osteoarthritis. Open Access Maced J Med Sci. 2018;6(3):493–497. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Sen EI, Eyigor S, Dikici Yagli M, et al. Effect of home-based exercise program on physical function and balance in older adults with sarcopenia: a multicenter randomized controlled study. J Aging Phys Act. 2021;29(6):1010–1017. [DOI] [PubMed] [Google Scholar]
- 50. Lazaridou A, Martel MO, Cornelius M, et al. The association between daily physical activity and pain among patients with knee osteoarthritis: the moderating role of pain catastrophizing. Pain Med. 2019;20(5):916–924. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are openly available at https://nda.nih.gov/oai.