Dispositional traits help explain individual differences in relationships between a radiographic knee osteoarthritis measure,pain,and physical function

Abstract

Background:

The concordance between radiograph-derived Kellgren–Lawrence (KL) scores for knee osteoarthritis (KOA) and experimental and clinical pain and KOA-related physical function is conflicting.

Objectives:

We investigate whether the inclusion of dispositional traits reduces variability between KOA radiographic findings, experimental pain, clinical pain, and function in individuals with knee pain.

Design:

This study is a cross-sectional, secondary analysis of data collected from the UPLOAD-II study.

Methods:

Adults aged 45–85 years with and without knee pain were enrolled. Data collected included sociodemographics, knee radiographs, experimental pain, clinical pain and function, and trait affect. Vulnerable and protective dispositional traits were classified from combined positive and negative trait affect measures. KL scores were determined from the knee radiographs. Unadjusted and adjusted (age, sex, comorbidities, and body mass index) regression analyses were completed with SAS version 9.4 (Cary, NC, USA).

Results:

The study included 218 individuals with a mean age of 58 years, 63.6% women, and 48.2% non-Hispanic black adults. Dispositional traits were associated with the experimental pain measures. No association between radiographic KOA and experimental pain was observed. In a combined and adjusted analysis, dispositional traits were predictive of knee punctate pain temporal summation (p = 0.0382). Both dispositional traits and radiographic KOA scores independently and combined were predictive of Graded Chronic Pain Scale pain and function, and Western Ontario and McMaster University pain and function (ps ⩽ 0.01). Improvements in R² were noted across all models with the inclusion of dispositional traits.

Conclusion:

Consideration of dispositional traits reduces the variability between radiographic KOA and pain and function. Non-pathological and associated pain-related psychological factors, dispositional traits might serve as parsimonious proxy tools to improve clinical assessments.

Registration:

N/A.

Plain language summary

Dispositional traits help explain individual differences in relationships between a radiographic knee osteoarthritis measure, pain, and physical function

Significance

• The concordance between radiographic knee osteoarthritis and experimental and clinical pain is conflicting.

• Dispositional traits comprise the infrastructure from which an individual interprets and interacts with the environment and are predictive of sensory sensitivity, response to stress, psychopathology, and behavior.

• Consideration of dispositional traits improves the congruence between knee osteoarthritis Kellgren-Lawrence scores, experimental pain, and clinical pain.

Keywords

dispositional traits function Kellgren–Lawrence scores knee osteoarthritis pain

Introduction

Knee osteoarthritis (KOA) is a common progressive joint disease characterized by pain and disability.¹ Radiographs are used to assess KOA severity. Kellgren–Lawrence (KL) scores are a common diagnostic classification system for evaluating grades of KOA with 0 showing no evidence of KOA to 4 indicating severe KOA.² Relationships between experimental and clinical pain and KOA radiographic classification based on KL scores are mixed. Individuals with KOA show heightened sensitivity compared to non-KOA controls on measures of experimental pain testing; however, across the different grades of KOA, the differences in sensitivity are less apparent.^3–5 Some research indicates that experimental pain sensitivity is more associated with clinical pain.^3,6 Findings between radiographic KOA and clinical pain are also mixed with several studies indicating no relationship and others showing significant relationships.^7–11 With approximately 30% of individuals with severe KOA being asymptomatic, the ability to identify the factors contributing to individual differences in pain and function would be a significant clinical benefit.¹²

Numerous psychological factors are associated with pain and function including anxiety, depression, catastrophizing, and fear avoidance. Less frequently considered are dispositional traits also known as temperament or personality traits which are neurobiological based and are associated with stress sensitivity and contribute toward cognitive, emotional, sensory, and behavioral functioning.^13–15 Dispositional traits exist across species and have been extensively studied in animal and clinical models.^13,16–20 Initial considerations of dispositional traits were described as early as 400 BC. Hippocrates is widely recognized for his identification of the four humors and their association with health and emotional functioning.²¹ Broadly investigated, numerous psychological theories and models have been developed around related constructs.^19,21–24 Preclinical and animal studies have provided further validating evidence for the underlying genetic and predisposing nature of dispositional traits.^16–18,25

Prior research has predominately focused on psychological pathologies in investigations of predictors of experimental and clinical pain. However, a strong body of evidence has been established in preclinical and clinical studies demonstrating the predictive utility of the non-pathological dispositional traits.^13,18,24 Although temperament is comprised of cognitive, sensory, physiological, and affective components, we and others have found a simple measure of trait affect significantly relevant.^26–32 Consistent with numerous models, Affect Balance Style (ABS) is a measure of combined levels of positive and negative trait affect.^27,33 We have shown ABS is associated with clinical pain, experimental pain, and pain-related psychological factors (e.g. somatic sensitivity, anxiety, catastrophizing, depression) in individuals with and without chronic pain.^26,27,29,33 In addition, aligning with preclinical models, we reported that the healthy and low ABS indicate more ‘protective’ traits compared to the more ‘vulnerable traits’ of the depressive and reactive ABS.²⁹ Dispositional traits may help inform clinical practice by accounting for some of the observed individual differences in radiographic KOA, pain, and function.

The purpose of the study was to investigate relationships between dispositional traits and radiographic KOA as measured by KL scores across measures of (1) experimental pain, (2) clinical pain and functioning, and whether the inclusion of dispositional traits with radiographic KOA scores improves the relationships with (1) experimental pain and (2) clinical pain and physical function. We hypothesized the relationships between radiographic KOA and experimental pain, clinical pain, and physical function would be limited. With consideration of dispositional traits, the variance accounted for in the study models would increase. Furthermore, in individuals with similar radiographic KOA classifications, those with vulnerable dispositional traits will report greater experimental and clinical pain and functional limitations compared to those with protective dispositional traits.

Methods

Participants

Adults between 45 and 85 years of age with and without chronic knee pain who self-identified as either non-Hispanic black (NHB) or non-Hispanic white (NHW) were recruited and enrolled at the University of Florida and the University of Alabama at Birmingham between August 2015 and May 2017.³⁴ Community advertisements and clinic referrals were the primary forms of recruitment. Exclusion criteria for the UPLOAD-2 study included: (1) cognitive impairment; (2) use of opioids on a daily basis; (3) hospitalization for a psychiatric illness in the preceding year; (4) a history of acute myocardial infarction, heart failure, or uncontrolled hypertension (BP > 150/95 mm Hg); (5) prosthetic knee replacements or other clinically significant surgery to the affected knee; (6) peripheral neuropathy; and/or (7) systemic diseases including rheumatoid arthritis, systemic lupus erythematosus, or fibromyalgia. This study was conducted in accordance with the Declaration of Helsinki. The UPLOAD-2 study was approved by the University of Florida Institution Review Board (IRB approval number 201400209) on 6 June 2014 and the University of Alabama at Birmingham Institution Review Board (IRB approval number 40915002) on 11 November 2014. All participants provided verbal and written informed consent prior to study participation. The current study consisted of data collected as part of the UPLOAD-2 study specific to participants with radiographic-derived KOA KL scores based on American College of Rheumatology criteria and a completed Positive and Negative Affect Schedule (PANAS).³⁵ This manuscript follows the STROBE checklist reporting guidelines.³⁶ An abstract of this data was presented at the United States Association for the Study of Pain conference in Cincinnati, Ohio in May 2022. Data are not publicly available but can be requested from the corresponding author. The current investigation is a cross-sectional analysis. The measures described are limited to those relevant to the identified research questions.

Procedures

Participants completed a health assessment and experimental pain testing session. Data collected included information specific to sociodemographics, health status, pain history, health outcomes, experimental pain testing, and knee radiographs. Questionnaires were completed and experimental pain testing was conducted specific to the index knee which was defined as the most painful knee for participants reporting knee pain and a randomized knee for the no knee pain comparison group. Participant questionnaires were reviewed for completeness to reduce bias.

Measures

Sociodemographic and health status

Baseline characteristics and health-related outcomes

Sociodemographic characteristics included age, sex, ethnicity/race, highest education completed, income range, and insurance status. Participants completed a health assessment and health history questionnaire including the report of current comorbidities including high blood pressure, heart disease, cancer, diabetes, asthma/breathing problems, kidney disease, thyroid problems, stroke, seizure, chronic pain, neurological disorder, depression, other mental health condition or other health problem. Height and weight were measured and body mass index (BMI) was calculated.

Clinical and radiographic criteria

Posterior–anterior and lateral radiographs of the knee were taken of the participant’s index knee. Posterior–anterior view radiographs were obtained using a Synaflexer with the X-ray beam centered on the joint line at a 10° caudal angulation. Lateral radiographs were obtained with the knee flexed at 40°. An experienced rheumatologist at the University of Florida, blinded to the knee group and participant characteristics, evaluated all radiographs for both study sites using the KL classification system (score range 0–4).³⁷

Experimental pain measures

Pressure pain threshold

Pressure pain threshold was assessed at the medial (n = 217) and lateral (n = 216) knee, using a handheld AlgoMed, Medoc digital pressure algometer. The pressure was increased at a constant rate of 30 kPa/s until the participant indicated the first sensation of pain, pain threshold. Pressure pain threshold testing was repeated three times at each site and the average at each site was used for analysis. The maximum pressure of the knee was 500 kPa. The order of testing sites was randomized.

Punctate pain temporal summation

Punctate stimuli were delivered with a calibrated 300g nylon monofilament on the index knee. First, a single stimulation was provided and then participants were asked whether the stimulation was painful and if so to provide a pain rating from 0 to 100. Then a series of 10 stimulations were given and participants were again asked if the stimulation was painful and if so to provide a pain rating from 0 to 100. This procedure was repeated twice. The average of the series stimulation was subtracted from the average of the single stimulation as a measure of punctate pain temporal summation and used for analysis (n = 218).

Clinical pain measures

Graded Chronic Pain Scale

The GCPS was used to assess the severity of knee pain and its impact on activities (n = 216). The measure is scored using two sub-scales, characteristic pain intensity (CPI) (0–100 score) and disability score (0–100 score) over a 6-month period. The higher the score, the higher the pain intensity and greater physical disability. The GCPS has demonstrated good internal consistency in prior research (α = 0.74).³⁸

Western Ontario and McMaster Universities Arthritis Index

The WOMAC was used to assess knee pain, function, and stiffness over the past 48 h.³⁹ The three subscales, pain (0–20 score), stiffness (0–8 score), and function (0–68 score), sum for a total WOMAC score (0–96 score) with higher scores indicating worse pain, stiffness, and functional limitations. The WOMAC pain (n = 217) and function (n = 218) were used in analyses. The WOMAC has demonstrated good internal consistency (α = 0.84–0.95) across several previous studies.^40,41

Total pain sites

Participants were asked to select from a list of pain sites if they experienced pain ‘more days than not over the past three months’. Bilateral body sites included hands, arms, shoulders, neck, head/face, chest, stomach, upper back, lower back, knees, legs (other than knees), or feet/ankles (0–24 sites). A greater number of pain sites is associated with worse health outcomes with three or more pain sites considered as widespread pain.^42,43

Dispositional traits

The PANAS is comprised of 10 positive and 10 negative words that reflect emotions and feelings.⁴⁴ Participants indicate on a five-point Likert scale how they feel ‘in general’ to each word from 1 = ‘very slightly or not at all’ to 5 = ‘extremely’. Scores range from 10 to 50 for both positive (PA) and negative (NA) affect, with higher scores indicating higher levels. The PANAS has a very good internal reliability with Cronbach alpha coefficient scores ranging from 0.86 to 0.90 for the PA scale and 0.84 to 0.87 for the NA scale.⁴⁵

Affect Balance Style (ABS) is determined based on published adult normative means for trait PANAS.³³ Although categorizing relationships between two independent continuous variables is not a recommended statistical approach, a review of foundational literature indicates the approach aligns with the conceptual models of several temperament theorists. The four group/quadrant classification of ABS styles is consistent with well-established lines of investigation that extend from Eysenck’s Personality Inventory,⁴⁶ to Watson, Clark, and Tellegen – the developers of the PANAS,³³ and persist in more current conceptual circumplex models of affect.¹⁹ The ABS groups were defined as follows: healthy ABS = high PA > 35 and low NA ⩽ 18.1; low ABS = low PA ⩽ 35 low NA ⩽ 18.1; reactive ABS = high PA > 35 high NA > 18.1, and depressive ABS = low positive PA ⩽ 35 and high NA > 18.^27,33 Protective and vulnerable dispositional traits were then categorized such that those individuals with healthy or low ABS were identified as having more protective traits and those individuals with reactive or depressive ABS were identified as having more vulnerable traits.²⁹ The two dispositional trait groups were used in study analyses (0 = protective, 1 = vulnerable) (n = 218).

Statistical analysis

Data were assessed for distributional form using visual inspection (histograms, boxplots) and normality was tested using the Shapiro–Wilk test. Missing variables were minimal (n = 2) and treated as missing. To test the research hypotheses of this paper, simple and multiple linear regression models were constructed. To assess the distributional assumptions of the regression models (normality of residuals, homogeneity of variance for residuals), residuals plots and normal probability plots were examined. We did not observe patterns within these graphs that would indicate a violation of the assumptions. Sample size calculations were completed for the parent study. Given this study is a secondary data analysis with a fixed sample size of 218, we have estimated and presented the effect sizes that would achieve 90% power (probability of being declared significant) for a sample size of 218 using a Type I error of 0.05. Additional explanatory variables included in all adjusted analyses were age, sex (ref = male), total number of comorbidities, and BMI. Participant characteristics are presented as descriptives.

To address question 1, investigate relationships between experimental pain, KL scores, and dispositional traits. Regression analyses were conducted for each outcome measure: pressure pain threshold at the medial knee, pressure pain threshold at the lateral knee, and punctate pain temporal summation as follows – Model 1: dispositional traits only, Model 2: KL scores only, Model 3: combined KL score and dispositional traits, Model 4: KL score, dispositional traits, and additional explanatory variables.

To address question 2, investigate relationships between clinical pain and functional measures, KL scores, and dispositional traits. Regression analyses were completed for each outcome measure GCPS CPI, WOMAC Pain, GCPS Interference, and WOMAC Function as follows – Model 1: dispositional traits only, Model 2: KL scores only, Model 3: combined KL score and dispositional traits, Model 4: KL score, dispositional traits, and additional explanatory variables.

Statistical analyses were completed using SAS version 9.4 (Cary, NC, USA). All statistical tests were conducted using a Type I error of 0.05.

Results

Descriptives

A total of 218 participants were included in the analysis. Participants were 57.4 ± 7.7 years of age, predominately female (65.1%) with 46.8% self-reporting as NHB (Table 1). There were no sex differences in dispositional traits or radiographic KOA as measured by KL scores. The NHB participants differed on socioenvironmental factors compared to NHW participants. NHB participants were younger, had less education, had lower income, and had a greater number of comorbidities compared to the NHW participants. Ethnicity/race is a self-reported construct where proportions of representation can be reported. However, as our ethnic/race groups do not match on relevant sociodemographic factors and the inclusion of covariates does not balance out group differences, and only a subgroup of each ethnic/race group is represented, a generalizable sample was not obtained.⁴⁷ The term sociodemographic groups rather than ethnic/race groups is a more accurate description and will be used for the group variable in the Results and Discussion.

Table 1.

Baseline characteristics of participants.

Variable	Total sample (n = 218)
Sociodemographic status
Age, M ± SD	57.4 ± 7.7
Gender, N (%)
Male	76 (34.9)
Female	142 (65.1)
Sociodemographic groups, N (%)
NHB	102 (46.8)
NHW	116 (53.2)
Site, N (%)
University of Florida	136 (62.4)
Univeristy Alabama at Birmingham	82 (37.6)
Education, N (%)
High school or less	127 (58.3)
Higher education	91 (41.7)
Income, N (%)
$0–29,999	109 (50.0)
$30,000–79,999	70 (32.1)
$80,000+	36 (16.5)
Not reported	3 (1.4)
Clinical characteristics
BMI, M ± SD	31.4 ± 7.5
No. comorbidities (0–14), N (%)
0	70 (32.1)
1–2	118 (54.1)
3+	30 (13.8)
KL scores, N (%)
2–4	68 (31.2)

BMI, body mass index; KL, Kellgren–Lawrence; M, mean; NHB, non-Hispanic black; NHW, non-Hispanic white; SD, standard deviation.

Comparisons were also completed within each of the dispositional trait groups, for example, for protective traits, the healthy and low ABS groups were compared, and for vulnerable traits the reactive and depressive ABS groups were compared specific to age, sex, number of pain sites, BMI, experimental and clinical pain, and functional measures. No significant differences were observed.

Relationships between dispositional traits, radiographic KOA, and experimental pain

Dispositional traits were significantly associated with medial (b = −57.0, p = 0.0047) and lateral (b = −60.2, p = 0.0065) pressure pain threshold and punctate temporal summation (b = 6.1, p = 0.0288) with the vulnerable traits group showing greater pain sensitivity (Table 2A, Model 1). Radiographic KOA was not associated with experimental pain (Table 2A, Model 2). In a regression analysis with KL score and dispositional traits, only dispositional traits were significantly associated with medial (b = −56.5, p = 0.0050) and lateral (b = −59.1, p = 0.0074) pressure pain threshold (Table 2A, Model 3). Inclusion of dispositional traits accounted for an additional 3% of the variance compared to KL scores alone.

Table 2.

Multivariate-multiple regression models of dispositional traits and KL scores by experimental pain, clinical pain, and function.

	(A) Experimental pain						(B) Clinical pain
	Pressure threshold medial knee		Pressure threshold lateral knee		Punctate temporal summation		GCPS CPI		GCPS interference		WOMAC pain		WOMAC function
Variables	b(SE)	t	b(SE)	t	b(SE)	t	b(SE)	t	b(SE)	t	b(SE)	t	b(SE)	t
Model 1
DT	−57.0(20.0)**	−2.7	−60.2(21.9)**	−2.8	6.0(2.8)*	2.2	18.5(4.1)***	4.5	19.6(5.0)***	3.9	3.8(0.7)***	5.1	12.2(2.5)***	4.9
R²	0.0296		0.0295		0.0246		0.0790		0.0680		0.1159		0.1080
Model 2
KL score	−3.8(7.1)	−0.5	−6.8(7.7)	−0.9	−0.7(0.7)	−1.0	5.5(1.3)***	4.3	6.7(1.5)***	4.4	0.9(0.2)***	4.1	3.9(0.8)***	5.2
R²	0.0015		0.0041		0.0037		0.0763		0.0820		0.0723		0.1213
ΔR²	−0.0146		−0.0254		−0.0209		−0.0027		0.0140		−0.0436		0.0133
Model 3
KL	−3.0(6.8)	−0.4	−5.8(7.5)	−0.8	−0.8(0.7)	−1.1	5.3(1.2)***	4.4	6.4(1.5)***	4.4	0.8(0.2)***	4.3	3.8(0.7)***	5.4
DT	−56.5(19.9)**	−2.8	−59.1(21.9)**	−2.7	6.2(2.8)	2.2	17.7(3.9)***	4.5	18.6(4.8)**	3.9	3.7(0.7)***	5.1	11.6(2.3)***	5.1
R²	0.0305		0.0324		0.0293		0.1482		0.1473		0.1795		0.2184
ΔR²	0.0290		0.0283		0.0256		0.0719		0.0653		0.1072		0.0971
Model 4
KL score	4.0(7.0)	0.6	−0.3(7.8)	0.0	−0.7(0.8)	−0.9	5.1(1.3)***	4.1	5.4(1.4)**	3.7	0.7(0.2)**	3.2	2.9(0.7)***	4.2
DT	−33.1(20.3)	−1.6	−41.5(23.3)	−1.8	6.8(2.8)*	2.4	11.3(3.9)**	2.9	12.2(4.7)**	2.6	2.7(0.7)**	3.8	8.9(2.3)**	3.9
Age	0.8(1.4)	0.6	0.8(1.4)	0.6	0.1(0.1)	1.0	−0.9(0.2)***	−3.9	−0.7(0.3)**	−2.7	−0.1(0.0)*	−2.6	−0.2(0.1)	−1.9
Sex	−91.8(20.3)***	−4.5	−47.2(21.9)*	−2.2	4.9(2.2)*	2.2	−8.0(3.7)	−2.2	−7.6(4.0)	−1.9	−1.2(0.6)	−1.9	−4.3(2.1)*	−2.1
Comorbidities	−11.6(7.4)	−1.6	−6.0(9.8)	−0.6	0.9(0.9)	1.0	5.6(1.4)***	4.0	8.0(1.5)***	5.2	1.1(0.2)***	4.7	3.7(0.7)***	5.1
BMI	−3.9(1.3)**	−2.9	−3.3(1.6)*	−2.1	−0.2(0.2)	−1.2	0.1(0.3)	0.2	0.3(0.3)	0.8	0.1(0.1)	1.4	0.3(0.2)	1.6
R²	0.2065		0.0976		0.0606		0.2527		0.2608		0.2941		0.3279
ΔR²	0.1760		0.0652		0.0313		0.1045		0.1135		0.1146		0.1095

p < 0.05. **p < 0.01. ***p < 0.0001.

BMI, body mass index; CPI, characteristic pain intensity; DT, dispositional traits (1 = protective and 2 = vulnerable); GCPS, graded chronic pain scale; KL score, Kellgren and Lawrence score; WOMAC, Western Ontario and McMaster University.

In combined and adjusted regression analyses including covariates age, sex, comorbidities, and BMI (Table 2A, Model 4), dispositional traits were significantly associated with punctate pain temporal summation (b = 6.8, p = 0.0166) while KL scores were not associated with any of the experimental pain measures. Female sex and higher BMI were associated with lower pressure pain threshold on both the medial and lateral knee. Female sex was also associated with greater punctate pain temporal summation.

Relationships between dispositional traits, radiographic KOA, and clinical pain and function

Dispositional traits were significantly associated with all measures of clinical pain (Table 2B, Model 1) with vulnerable traits related to worse pain and lower function. Radiographic KOA was also associated with all clinical pain and functional measures such that higher KOA was associated with worse pain and lower function (Table 2B, Model 2). In a regression analysis with KL score and dispositional traits, both KL scores and dispositional traits were associated with all clinical pain and functional measures (ps < 0.01) (Table 2B, Model 3). Inclusion of dispositional traits accounted for an additional 6–10% of the variance compared to KL scores alone. Figure 1 is a residual plot of clinical pain and function measures by KL scores with a trend line for dispositional traits. Individuals with vulnerable traits reported greater clinical pain and interference on all clinical measures (p < 0.001).

In adjusted regression analyses including age, sex, comorbidities, and BMI (Table 2B, Model 4), KL scores and dispositional traits were significantly associated with all clinical pain and functional measures (ps < 0.01). A greater number of comorbidities was associated with all outcomes while higher age was associated with lower GCPS CPI, GCPS interference, and WOMAC pain. Male sex was associated with greater GCPS CPI and greater WOMAC function.

Discussion

Our study investigated the relationships between experimental pain, clinical pain, physical functioning, and radiographic KOA measured by dispositional traits and KL scores. Dispositional traits were associated with all experimental pain, clinical pain, and functional measures. Radiographic KOA was not associated with experimental pain but was associated with all clinical pain and functional measures. As hypothesized, inclusion of dispositional traits in models with radiographic KOA, variability was reduced and all models were strengthened. Dispositional traits help to explain some of the observed individual differences between radiographic KOA, experimental pain, clinical pain, and physical function. Dispositional traits are non-pathological, neurobiological-based, and are associated with pain-related psychological measures. As such, a brief measure of dispositional traits may serve as a clinical tool to improve the assessment and treatment planning process for individuals with chronic pain.

Relationships between dispositional traits, radiographic KOA, and experimental pain

Relationships between experimental pain measures and dispositional traits were indicated. In general, as anticipated, individuals with vulnerable dispositional traits had lower pressure pain thresholds and greater punctate pain temporal summation compared to individuals with protective dispositional traits. In adjusted analyses, only punctate pain temporal summation was associated with dispositional traits. One factor that may contribute to a limited indication of relationships between dispositional traits and experimental pain is a brief pain stimuli exposure period. Our previous investigation of ABS in healthy subjects indicated associations were limited to the ischemic measure, which is a prolonged painful stimulus.³³

Contrary to previous research, we found no significant relationships between KL scores and the experimental pain measures. Both pressure pain threshold and punctate pain temporal summation have been associated with KOA.³ There is some evidence to suggest that radiographic evaluations of KOA such as KL scores may mis-estimate disease severity.^48,49 Our study sample was comprised of individuals who reported knee pain in the prior month with over 68% not meeting radiographic KOA criteria (KL score ⩾2). In addition, the sample sizes were small for those individuals with a KL score of 2 or greater. Previous research indicated greater KL scores were associated with pressure pain threshold likely related to mechanical hyperalgesia of the knee.⁵⁰ Given that a majority of our participants did not have KL scores of 3 or 4, the small representation of individuals meeting KOA criteria may explain the lack of relationship. Importantly, the inclusion of dispositional traits increased the variance accounted for in all the experimental pain models.

Relationships between dispositional traits, radiographic KOA, and clinical pain and function

Dispositional traits were associated with clinical pain measures. The pattern of findings aligns with the heightened sensory sensitivity in individuals with vulnerable traits compared to those with protective traits.^31,33,51 Most striking is that a review of the means between the protective and vulnerable dispositional traits for GCPS CPI indicates an approximate average difference of 18.5 on a numerical rating scale of 0–100 between the same KL scores, a recognized clinically significant difference.⁵² Our findings help explain what has been a perplexing clinical observation, significant individual differences in pain experiences within individuals with similar stages of radiographic KOA. The pain intensity report for individuals with a KL of 0 and vulnerable dispositional traits is similar to the pain intensity report of individuals with a KL score of 4 and protective dispositional traits.

Radiological KOA was also significantly associated will all clinical pain measures in unadjusted and adjusted analyses matching what was found in some previous studies.^7,8,50,53 As we had a significant proportion of participants reporting knee pain consistent with American College of Rheumatology criteria KL score of 0 and 1, the greater representation of individuals with no radiographic KOA may contribute to the observed relationships between KL scores and clinical pain compared to some of the prior publications that did not find a relationship.⁵⁴ Inclusion of dispositional traits improved the predictive modeling for all of the clinical pain measures, with overall models accounting for 25–33% of the variance in clinical pain experiences. In addition, as socioenvironmental factors are associated with a greater prevalence of arthritis, consideration of social-environmental factors may further inform findings regarding the relationships between radiographic KOA, dispositional traits, pain, and functioning.^49,55

Strengths, limitations, and future directions

There are several strengths in the current investigation. First, the study includes a large sample of community-dwelling adults with a representation of NHB and NHW from differing socioenvironmental backgrounds. Second, all radiographs were read by a rheumatologist blinded to the participant characteristics (sex, age, sociodemographic group). Third, we included measures capturing both recent, that is, past 48 h (WOMAC), and within a 6-month timeframe (GCPS), to assess pain and function.

Some limitations warrant acknowledgment. First, the study is cross-sectional in design. Evaluating the prospective relationships between dispositional traits, radiographic KOA, experimental pain, clinical pain, and physical function would be highly informative. Second, 68.8% of our sample (n = 150) had KL scores < 2; thus, they did not meet the radiographic criteria for KOA. However, only 11% (n = 23) reported not having clinical pain or interference/disability. Thus, pain and functional limitations reported are possibly indicative of prodromal or clinical and not radiographic features of KOA or due to other causes.^35,49 In addition, as more than half of the participants did not meet KOA radiographic criteria, the generalizability of findings to a KOA population is limited. Third, Figure 1 represents a trend line as the sample size for the KL score stratified by dispositional traits. Our KL scores had limited sample sizes for KL 2–4 (with a total of n = 68) with KL 3 having the smallest sample size (n = 16). Fourth, although the PANAS questionnaire is a brief instrument that can be implemented in a clinical setting, it is limited to a focus on trait affect which is only one domain of dispositional trait functioning. Future investigations of a brief and more comprehensive measure of dispositional traits would be useful.²⁹ Fifth, arthritis prevalence is associated with socioenvironmental factors, future investigations warrant consideration of these variables.⁵⁵ Finally, an adjustment in the statistical modeling approach might also be more informative.⁵⁶ The estimates from our regression models should be interpreted as the estimated effect of the variable upon the outcome after adjustment for the other predictors included in the models. In the likely event that another predictor exists that has not been included in our models, is correlated with the outcome of our model, and is correlated with a subset of the predictors in our model, then the estimates within our tables may be statistically biased.

Figure 1.

Clinical pain and function by Kellgren-Lawrence Scores with consideration for dispositional traits.

Dispositional traits serve as the interpretative infrastructure between an animal or human’s stress response system and the environment.^{14,15,21,57,58} Rather than serving as an independent predictor, it is the combination of stress exposures with dispositional traits that increases or decreases health risks. Children and adults with vulnerable traits confronted by challenges without adaptive coping mechanisms are at greater risk for poor outcomes.^{19,20,59–62} Importantly, preclinical and clinical models indicate that interventions can improve function, modifying predisposing dispositional vulnerabilities.^25,57,58 Importantly, dispositional traits are non-pathological, classifiable across all individuals, and highly associated with a list of commonly assessed pain-related psychological measures (depression, anxiety, catastrophizing, fear avoidance).^26,29,33 A measure of dispositional traits might serve as a parsimonious tool to improve pain-related assessments and might have utility in individualized treatment planning.

Conclusion

Poor concordance between radiographic KOA and patient reports of pain and functioning complicates treatment planning and creates challenges for clinicians and researchers. Consideration of dispositional traits improved the congruence between KOA measured by KL scores and pain and function in mid to older adults with or at risk for knee osteoarthritis. Findings indicate a consistent heightened sensitivity to pain in individuals with vulnerable traits compared to those with protective traits. Future research assessing whether the assessment of dispositional traits might serve as a tool to improve understanding of individual differences in pain-related experiences.

Supplemental Material

sj-docx-1-tab-10.1177_1759720X241235805 – Supplemental material for Dispositional traits help explain individual differences in relationships between a radiographic knee osteoarthritis measure, pain, and physical function

Supplemental material, sj-docx-1-tab-10.1177_1759720X241235805 for Dispositional traits help explain individual differences in relationships between a radiographic knee osteoarthritis measure, pain, and physical function by Angela M. Mickle, Roland Staud, Cynthia S. Garvan, Daniel A. Kusko, Nicola Sambuco, Brittany R. Addison, Kevin R. Vincent, David T. Redden, Burel R. Goodin, Roger B. Fillingim and Kimberly T. Sibille in Therapeutic Advances in Musculoskeletal Disease

Footnotes

Acknowledgements

The authors would like to thank the UPLOAD-2 participants, research teams and Clinical and Translational Science Institute nursing staff and processing labs at the University of Florida and the University of Alabama at Birmingham.

Declarations

ORCID iD

Angela M Mickle

Supplemental material

Supplemental material for this article is available online.

References

Cui

Wang

, et al. Global, regional prevalence, incidence and risk factors of knee osteoarthritis in population-based studies. EClinicalMedicine 2020; 29–30: 100587.

Kohn

Sassoon

Fernando

. Classifications in brief: Kellgren-Lawrence classification of osteoarthritis. Clin Orthop Relat Res 2016; 474: 1886–1893.

Arant

Katz

Neogi

. Quantitative sensory testing: identifying pain characteristics in patients with osteoarthritis. Osteoarthr Cartil 2022; 30: 17–31.

Frey-Law

Bohr

Sluka

, et al. Pain sensitivity profiles in patients with advanced knee osteoarthritis. Pain 2016; 157: 1988–1999.

Kavchak

AJE

Fernández-de-las-Peñas

Rubin

, et al. Association between altered somatosensation, pain, and knee stability in patients with severe knee osteoarthrosis. Clin J Pain 2012; 28: 589–594.

Suokas

Walsh

McWilliams

, et al. Quantitative sensory testing in painful osteoarthritis: a systematic review and meta-analysis. Osteoarthr Cartil 2012; 20: 1075–1085.

Rajpoot

Peepra

Pandey

, et al. Clinico-radiological correlation of osteoarthritis knee using Western Ontario and McMaster Universities score and Kellegren and Lawrance grading. IJRMS 2016: 4873–4876.

Wang

Kim

Felson

, et al. Radiographic knee osteoarthritis and knee pain: cross-sectional study from five different racial/ethnic populations. Sci Rep 2018; 8: 1364.

Cubukcu

Sarsan

Alkan

. Relationships between pain, function and radiographic findings in osteoarthritis of the knee: a cross-sectional study. Arthritis 2012; 2012: 984060.

10.

Dieppe

Cushnaghan

Shepstone

. The Bristol ‘OA500’ study: progression of osteoarthritis (OA) over 3 years and the relationship between clinical and radiographic changes at the knee joint. Osteoarthr Cartil 1997; 5: 87–97.

11.

Litwic

Edwards

Dennison

, et al. Epidemiology and burden of osteoarthritis. Br Med Bull 2013; 105: 185–199.

12.

Son

Hong

Kim

D-H

, et al. Absence of pain in subjects with advanced radiographic knee osteoarthritis. BMC Musculoskelet Disord 2020; 21: 640.

13.

Caspi

Harrington

Milne

, et al. Children’s behavioral styles at age 3 are linked to their adult personality traits at age 26. J Pers 2003; 71: 495–513.

14.

Robinson

. How brain arousal systems determine different temperament types and the major dimensions of personality. Pers Individ Dif 2001; 31: 1233–1259.

15.

Strelau

. The regulative theory of temperament: current status. Pers Individ Dif 1996; 20: 131–142.

16.

Lages

McNaughton

. Non-human contributions to personality neuroscience - from fish through primates. An introduction to the special issue. Personal Neurosci 2022; 5: e11.

17.

Powell

Horvath

Gilliam

, et al. Behavioral and physiological reactions to a sudden novel object in the weanling horse: quantitative phenotypes for future GWAS. Genes 2023; 14: 593.

18.

Fox

Harris

Rosso

, et al. Infant inhibited temperament in primates predicts adult behavior, is heritable, and is associated with anxiety-relevant genetic variation. Mol Psychiatry 2021; 26: 6609–6618.

19.

Posner

Russell

Peterson

. The circumplex model of affect: an integrative approach to affective neuroscience, cognitive development, and psychopathology. Dev Psychopathol 2005; 17: 715–734.

20.

Rettew

McKee

. Temperament and its role in developmental psychopathology. Harv Rev Psychiatry 2005; 13: 14–27.

21.

Rothbart

Ahadi

Evans

. Temperament and personality: origins and outcomes. J Pers Soc Psychol 2000; 78: 122–135.

22.

Watson

Tellegen

. Toward a consensual structure of mood. Psychol Bull 1985; 98: 219–235.

23.

Chess

. Studies in temperament: a paradigm in psychosocial research. Yale J Biol Med 1990; 63: 313–324.

24.

Klein

Gaspardo

Martinez

, et al. Pain and distress reactivity and recovery as early predictors of temperament in toddlers born preterm. Early Hum Dev 2009; 85: 569–576.

25.

Stead

Clinton

Neal

, et al. Selective breeding for divergence in novelty-seeking traits: heritability and enrichment in spontaneous anxiety-related behaviors. Behav Genet 2006; 36: 697–712.

26.

Hassett

Goesling

Mathur

, et al. Affect and low back pain: more to consider than the influence of negative affect alone. Clin J Pain 2016; 32: 907–914.

27.

Hassett

Simonelli

Radvanski

, et al. The relationship between affect balance style and clinical outcomes in fibromyalgia. Arthritis Rheum 2008; 59: 833–840.

28.

Toussaint

Vincent

McAllister

, et al. A comparison of fibromyalgia symptoms in patients with healthy versus depressive, low and reactive affect balance styles. Scand J Pain 2014; 5: 161–166.

29.

Sambuco

Mickle

Garvan

, et al. Vulnerable dispositional traits and chronic pain: predisposing but not predetermining. J Pain 2022; 23: 693–705.

30.

Badil Güloğlu

Tunç

. The assessment of affective temperament and life quality in myofascial pain syndrome patients. Int J Clin Pract 2022; 26: 79–84.

31.

Favaretto

Gögele

Bedani

, et al. Pain sensitivity is modulated by affective temperament: results from the population-based CHRIS Affective Disorder (CHRIS-AD) study. J Affect Disord 2022; 316: 209–216.

32.

Knefel

Zeilinger

Erfurth

, et al. Affective temperament, fatigue, and pain in cancer patients. J Affect Disord 2023; 340: 80–87.

33.

Sibille

Kindler

Glover

, et al. Affect balance style, experimental pain sensitivity, and pain-related responses. Clin J Pain 2012; 28: 410–417.

34.

Roux

Saraux

Mazieres

, et al. Screening for hip and knee osteoarthritis in the general population: predictive value of a questionnaire and prevalence estimates. Ann Rheum Dis 2008; 67: 1406–1411.

35.

Altman

Asch

Bloch

, et al. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum 1986; 29: 1039–1049.

36.

Cuschieri

. The STROBE guidelines. Saudi J Anaesth 2019; 13: S31–S34.

37.

Kellgren

Lawrence

. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957; 16: 494–502.

38.

Von Korff

Ormel

Keefe

, et al. Grading the severity of chronic pain. Pain 1992; 50: 133–149.

39.

Thumboo

Chew

Soh

. Validation of the Western Ontario and Mcmaster University osteoarthritis index in Asians with osteoarthritis in Singapore. Osteoarthr Cartil 2001; 9: 440–446.

40.

Bruce

Fries

. Longitudinal comparison of the Health Assessment Questionnaire (HAQ) and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC). Arthritis Care Res 2004; 51: 730–737.

41.

Stratford

Kennedy

Woodhouse

, et al. Measurement properties of the WOMAC LK 3.1 pain scale. Osteoarthr Cartil 2007; 15: 266–272.

42.

Mose

Kent

Smith

, et al. Number of musculoskeletal pain sites leads to increased long-term healthcare contacts and healthcare related costs – a Danish population-based cohort study. BMC Health Serv Res 2021; 21: 980.

43.

Wolfe

Butler

Fitzcharles

, et al. Revised chronic widespread pain criteria: development from and integration with fibromyalgia criteria. Scand J Pain 2019; 20: 77–86.

44.

Watson

Clark

Tellegen

. Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol 1988; 54: 1063–1070.

45.

Magyar-Moe

. Chapter 3 - Positive Psychological Tests and Measures. In: Magyar-Moe

(ed) Therapist’s Guide to Positive Psychological Interventions. San Diego: Academic Press, 2009, pp. 43–72.

46.

Eysenck

SBG

. Manual of the Eysenck Personality Inventory. London: University of London Press, 1964.

47.

Miller

Chapman

. Misunderstanding analysis of covariance. Journal of Abnormal Psychology 2001; 110: 40–48.

48.

Abdelaziz

Balde

Citak

, et al. Kellgren–Lawrence scoring system underestimates cartilage damage when indicating TKA: preoperative radiograph versus intraoperative photograph. Arch Orthop Trauma Surg 2019; 139: 1287–1292.

49.

Pierson

Cutler

Leskovec

, et al. An algorithmic approach to reducing unexplained pain disparities in underserved populations. Nat Med 2021; 27: 136–140.

50.

Satake

Izumi

Aso

, et al. Comparison of predisposing factors between pain on walking and pain at rest in patients with knee osteoarthritis. J Pain Res 2021; 14: 1113–1118.

51.

Aho

Sipilä

Kalso

, et al. Temperament and character dimensions differ in chronic post-surgical neuropathic pain and cold pressure pain. Scand J Pain 2022; 22: 515–525.

52.

Farrar

Portenoy

Berlin

, et al. Defining the clinically important difference in pain outcome measures. Pain 2000; 88: 287–294.

53.

Tavares

DRB

Moça Trevisani

Frazao Okazaki

, et al. Risk factors of pain, physical function, and health-related quality of life in elderly people with knee osteoarthritis: a cross-sectional study. Heliyon 2020; 6: e05723.

54.

Bedson

Croft

. The discordance between clinical and radiographic knee osteoarthritis: a systematic search and summary of the literature. BMC Musculoskeletal Disorders 2008; 9: 116.

55.

Rethorn

Cook

, et al. Association of burden and prevalence of arthritis with disparities in social risk factors, findings from 17 US States. Prev Chronic Dis 2022; 19: E08.

56.

Mickle

Tanner

Holmes

, et al. Vulnerable dispositional traits and poverty are associated with older brain age. J Pain, Abstract 2023; 24.

57.

Lansade

Bouissou

Boivin

. Temperament in preweanling horses: development of reactions to humans and novelty, and startle responses. Dev Psychobiol 2007; 49: 501–513.

58.

Shackman

Tromp

DPM

Stockbridge

, et al. Dispositional negativity: an integrative psychological and neurobiological perspective. Psychol Bull 2016; 142: 1275–1314.

59.

Hur

Stockbridge

Fox

, et al. Dispositional negativity, cognition, and anxiety disorders: an integrative translational neuroscience framework. Prog Brain Res 2019; 247: 375–436.

60.

Johansson

Guo

Duberstein

, et al. Midlife personality and risk of Alzheimer disease and distress: a 38-year follow-up. Neurology 2014; 83: 1538–1544.

61.

Aschenbrenner

Petros

McDade

, et al. Relationships between big-five personality factors and Alzheimer’s disease pathology in autosomal dominant Alzheimer’s disease. DADM 2020; 12: e12038.

62.

Terracciano

Bilgel

Aschwanden

, et al. Personality associations with amyloid and tau: results from the baltimore longitudinal study of aging and meta-analysis. Biol Psychiatry 2022; 91: 359–369.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB