Supplemental Digital Content is Available in the Text.
Approximately one in 5 patients experienced high-impact chronic pain after total joint arthroplasty. Several modifiable predictors highlight opportunities for presurgical optimization.
Keywords: High-impact chronic pain, Total joint arthroplasty, Chronic pain, Postoperative outcomes, Outcomes
Abstract
Introduction:
High-impact chronic pain (HICP) is associated with impaired function, diminished quality of life, and greater healthcare utilization. Although total joint arthroplasty (TJA) is effective for osteoarthritis, up to 20% of patients report persistent pain and limitations.
Objective:
To examine transitions in HICP status and identify predictors of postoperative HICP following TJA.
Methods:
This was a retrospective longitudinal cohort study of 6631 patients who underwent total knee (46.6%), hip (31.2%), or shoulder (22.2%) arthroplasty at a large US academic medical center between January 2018 and August 2024. Transitions in HICP status were described, and multivariable logistic regression was used to evaluate predictors of postoperative HICP. Results were reported as odds ratios (OR) with 95% confidence intervals (CI). Model performance was assessed using classification accuracy, receiver operating characteristic analysis, and the area under the curve.
Results:
Preoperatively, 45.3% of patients reported HICP, which declined to 15.2% after surgery. However, 4.8% developed new-onset HICP, yielding an overall postoperative prevalence of 20.0%. Preoperative HICP (OR = 2.50, 95% CI: 2.10–2.94), lower physical function (OR = 0.95, 95% CI: 0.93–0.96), higher pain intensity (OR = 1.07, 95% CI: 1.04–1.10), and postoperative opioid exposure (OR = 1.83, 95% CI: 1.59–2.11) were the strongest predictors. Additional factors included younger age, sex, alcohol use, revision surgery, and comorbidity burden.
Conclusion:
Although TJA substantially reduces HICP prevalence, 1 in 5 patients continues to experience significant postoperative pain. Early identification of at-risk individuals and the integration of targeted perioperative interventions are critical to optimize recovery and long-term outcomes.
1. Introduction
High-impact chronic pain (HICP) is prevalent in the United States, with approximately 21% of US respondents reporting persistent chronic pain, and more than 60% of those respondents continued to have chronic pain in the following year.31 On an individual level, the consequences of HICP include significant disruptions in daily activities and social roles, psychological distress, continued healthcare utilization, and long-term opioid usage.4,8,23,35,36,51 Because of the prevalence and impact, HICP has emerged as a public health concern.23,40
In total joint arthroplasty (TJA), 15% to 30% of patients report chronic postoperative pain17,52 and dissatisfaction following TJA,19 despite improvements in physical function and quality of life for many individuals.13 Although OA is the predominant indication for TJA, the inclusion of patients undergoing surgery for biomechanical dysfunction, anatomical deformity, fracture, or trauma can contribute to variability in patient pain profiles.5 Total joint arthroplasty procedures address the structural and mechanical sources of joint pain by restoring alignment, stability, and load transmission3,32; however, it does not necessarily resolve the underlying neurobiological and psychological factors that contribute to chronic pain.48,49 Accordingly, a 2022 cross-sectional study reported that approximately 10% of patients undergoing TJA reported having postoperative HICP, with rates (95% confidence intervals [CI]) reported for knee (9.8%–13.3%), hip (8.3%–11.8%), and shoulder (7.6%–16.3%) arthroplasty varying slightly.17 Thus, chronic postsurgical pain in this population, including HICP, is not solely a biomechanical problem but a multifactorial condition requiring a broader biopsychosocial framework.14,29
Two significant gaps in the literature related to HICP and TJA remain. First, robust estimates of HICP have primarily been limited to population-based, cross-sectional cohorts without indication of specific intervention or diagnostic subgroup.31,34 Although longitudinal assessment of pain outcomes in TJA has been evaluated, the evaluation of HICP classification in TJA has not yet been assessed in a large, longitudinal cohort. As a result, risk factors associated with the development of postoperative HICP in these populations have not been clearly identified. This study aimed to address gaps in the literature by building on previous findings16,32,35 and recent reviews14,29 that have identified potential risk factors for postoperative HICP in TJA patients. Specifically, our aims were (1) to characterize the transition to and from postoperative HICP status in a cohort of patients undergoing hip, knee, or shoulder TJA and (2) to identify sociodemographic, preoperative, perioperative, and patient-reported outcome factors associated with postoperative HICP. The overarching goal was to identify risk factors, both modifiable and non-modifiable, associated with postoperative HICP with the intent of identifying at-risk patients to inform targeted interventions, optimize clinical protocols, and improve postoperative outcomes.
2. Methods
2.1. Study setting and participants
This was an institutional review board–approved (Pro00091740) retrospective cohort study of patients who underwent total hip arthroplasty, total knee arthroplasty, or total shoulder arthroplasty within a high-volume department at an academic center in the United States between January 24, 2018, and August 16, 2024. Inclusion criteria for the study were patients who underwent a primary or revision arthroplasty procedure (Supplemental Table 1, http://links.lww.com/PR9/A385). Primary exclusion criteria were incomplete follow-up Patient-Reported Outcomes Measurement Information Systems (PROMIS) Pain Interference (PI) measure, were younger than 18 years, died within 1 year of surgery, or had missing/incomplete data fields for surgical type or joint (Fig. 1). Patients and members of the public were not directly involved in the design, conduct, or analysis of this observational study. However, the study was informed by priorities identified in the existing literature and clinical practice guidelines to ensure relevance to patient populations.9 Expanded operational definitions of the variables are provided in Supplemental Table 2, http://links.lww.com/PR9/A385.
Figure 1.
Flow diagram of the cohort, inclusion, and exclusion criteria.
2.2. Variables
2.2.1. Demographic factors
Demographic data were extracted from the electronic health record (EHR) relative to the TJA surgery. These variables included age at surgery, sex, body mass index (BMI) at surgery, race, ethnicity, and the area deprivation index (ADI).37
2.2.2. Preoperative risk factors
We recorded comorbidities, including the 30 comorbidities in the van Walraven Elixhauser Comorbidity Index,50 the American Society of Anesthesiologists Classification (ASA),11,15 diagnosis of anxiety or depression, and self-reported tobacco and alcohol use. These comorbidities were considered present if documented in the EHR within 12 months of the surgical encounter.
2.2.3. Perioperative factors
Data regarding the surgical event were recorded from the EHR. These included the patient's location of discharge after surgery, length of stay in days (LOS) in the hospital after surgery, the TJA site (hip, knee, or shoulder), and the surgery type (primary or revision TJA). Opioid exposure as a covariate was defined as the presence or absence of a completed prescription recorded within the EHR for a medication classified as an opioid analgesic. We implemented a washout period of 10 days, as it is expected that most patients will have acute pain management with an opioid postoperatively, and this washout period aligns with the mandated prescribing practices in North Carolina and postoperative care.7,26
2.2.4. Patient-reported outcomes
As part of standard care, patients are administered PROMIS measures PI and physical function and the numerical pain rating scale at orthopedic healthcare encounters within the health system and recorded in the EHR (Epic Systems).20 For this study, we extracted PROMIS PI, PROMIS physical function, and numerical pain rating scale scores pre- and 6 months postoperatively, aligning with the HICP definition and expected recovery trajectories in TJA.9,39 If multiple scores were recorded for a patient, the score closest to the surgical event (preop) and the score at 180 days (postop) were extracted.
2.2.5. Determining high-impact chronic pain status
High-impact chronic pain is defined as the presence of pain on at least half of the days in the previous 3 to 6 months with substantial restriction of functional participation in work, social, and self-care activities.12 The assessment of HICP recommended by the US Federal Pain Research Strategy is a core measure in the minimum data for the NIH Task Force on chronic low back pain and has been endorsed by the VA as a core measure.10,28 Despite these recommendations, HICP status is not an outcome routinely assessed in clinical care, likely due to the administrative burden of additional questionnaire items. To estimate HICP status for patients undergoing TJA without increasing burden, we have reported that PROMIS-PI accurately estimates HICP status in these patients.18 However, these analyses did not differentiate other chronic pain grades (eg, Bothersome Chronic Pain) beyond HICP status. Although the PROMIS-PI measure does not directly assess pain frequency or duration, it does evaluate the consequences of pain across relevant aspects of one's life, including its impact on social, cognitive, emotional, physical, and recreational engagement.2 Thus, it was highly associated with HICP status in our prior analyses and is a pragmatic way to estimate HICP status without increasing patient response burden. PROMIS PI is scored on a T-score metric, where 50 is the mean and 10 is the standard deviation of the US general population. A higher score on the PROMIS PI reflects greater impairment and a lower score reflects less impairment. Using cutoffs identified in our prior work, we defined HICP status as a PROMIS-PI score of 65 or higher for both pre- and postoperative statuses. Briefly, this cutoff score was derived from psychometric analyses of a cohort of TJA patients completing PROMIS measures and the Graded Chronic Pain Scale Revised. These analyses indicated that PROMIS-PI was the better measure, with a cutoff score of 65 or higher indicating at least a 96% probability of HICP status for TJA at the hip, knee, and shoulder.18
2.3. Data analysis and statistical considerations
Cohort characteristics, including pre- and postoperative HICP rates, were summarized as frequencies and percentages for categorical variables and means and standard deviations for continuous variables. To address aim 2, we fit univariable and multivariable logistic regression models to assess the risk of postoperative HICP. Before fitting models, we verified the linearity assumptions for continuous variables and identified and removed any outliers (n = 1). We examined the extent of missing data among candidate predictor variables and imputed missing values for variables considered essential to the analysis using multiple imputation methods. Twenty imputed datasets were created using linear regression, logistic regression, or discriminant analysis to generate possible values based on the observed data format. Missing data included BMI (n = 725, 11%), ASA score (n = 605, 9%), national ADI rank (n = 1776, 27%), anxiety (n = 1610, 24%), depression (n = 1610, 24%), tobacco use (n = 808, 12%), alcohol use (n = 835, 13%), pain intensity (n = 2141, 32%), and hospital length of stay (n = 4, <1%).
Next, we evaluated a set of candidate variables to identify predictors of postoperative HICP. These predictors were selected based on our prior cross-sectional study and recent scoping reviews.14,17,29 They included demographic characteristics (age, sex, race, ethnicity, and other sociodemographic factors), as well as preoperative clinical risk factors such as BMI, ASA score, Elixhauser comorbidity index, presence of anxiety or depression, tobacco or alcohol use, and patient-reported outcomes including PROMIS PI, PROMIS Physical Function, and pain intensity. Perioperative factors included joint type, surgical procedure, discharge location, length of hospital stay, and opioid exposure. All candidate predictors were obtained from the EHR. First, we performed individual univariable logistic regression models with postoperative HICP (Yes) as the outcome and each candidate predictor, using each of the 20 imputed datasets. The resulting parameter estimates and standard errors for each model were then combined using Rubin rule. The linearity assumption for continuous variables was evaluated using restricted cubic splines, which indicated that age and BMI should be modeled using a linear spline.
Variables found associated (P < 0.20) in univariable comparisons (all but ethnicity) were included in the multivariable logistic regression model applied to each imputed dataset. Collinearity was not detected among the covariates using variance inflation factors. The parameter estimates from each model were combined using Rubin rule, and the final model results were presented as odds ratios (ORs) with 95% CIs. We evaluated the performance of the multivariable model by examining the correct classification rate (overall accuracy) and receiver operating characteristic (ROC) analysis (sensitivity, specificity, negative and positive predictive value using 0.50 as the threshold probability value), and the area under the ROC curve (AUC) with 95% CI. All analyses were performed using SAS version 9.4 (SAS Institute, Inc., Cary, NC).
3. Results
The cohort of patients in this analysis included 6631 patients who underwent TJA. Patients were an average of 66.1 year old (SD 10.1), predominantly female (59.8%), and Caucasian (76.5%). Most patients in the cohort underwent primary TJA (91.0%) compared with revision surgery (9.0%). Consistent with the expected surgical volume, the most frequent procedures were total knee arthroplasty (46.6%), followed by THA (31.2%) and total shoulder arthroplasty (22.2%). The average Elixhauser sum score was 3.1 (SD 2.3) for comorbidities, and most patients were discharged home (86.8%) (Table 1). The patient's pre- and postoperative raw scores on the outcome measure are shown in Table 2.
Table 1.
Cohort demographics by postoperative HICP status.
| Variables | Postoperative HICP status | Total (N = 6631) | Missing | |
|---|---|---|---|---|
| Yes (N = 1324) | No (N = 5307) | |||
| Age at surgery, y | 64.4 (11.0) | 66.5 (9.8) | 66.1 (10.1) | 0 |
| Female sex | 816 (61.6%) | 3150 (59.4%) | 3966 (59.8%) | 0 |
| Race | 0 | |||
| White/Caucasian | 907 (68.5%) | 4164 (78.5%) | 5071 (76.5%) | |
| Black/African American | 344 (26.0%) | 917 (17.3%) | 1261 (19.0%) | |
| Asian | 16 (1.2%) | 94 (1.8%) | 110 (1.7%) | |
| Native American/Pacific Islander/Alaskan Native | 12 (0.9%) | 24 (0.5%) | 36 (0.5%) | |
| Other | 23 (1.7%) | 46 (0.9%) | 69 (1.0%) | |
| Not reported | 22 (1.7%) | 62 (1.2%) | 84 (1.3%) | |
| Ethnicity | 0 | |||
| Not Hispanic/Latino | 1267 (95.7%) | 5071 (95.6%) | 6338 (95.6%) | |
| Hispanic/Latino | 32 (2.4%) | 98 (1.8%) | 130 (2.0%) | |
| Not reported | 25 (1.9%) | 138 (2.6%) | 163 (2.5%) | |
| National ADI quartiles | 1776 | |||
| First: 0 < ADI <24 | 215 (21.7%) | 1112 (28.8%) | 1327 (27.3%) | |
| Second: 24 ≤ ADI <40 | 247 (25.0%) | 1130 (29.2%) | 1377 (28.4%) | |
| Third: 40 ≤ ADI <61 | 261 (26.4%) | 894 (23.1%) | 1155 (23.8%) | |
| Fourth: ADI ≥ 61 | 267 (27.0%) | 729 (18.9%) | 996 (20.5%) | |
| Elixhauser summary score | 3.8 (2.5) | 2.9 (2.2) | 3.1 (2.3) | 0 |
| ASA | 605 | |||
| Healthy | 8 (0.7%) | 66 (1.4%) | 74 (1.2%) | |
| Mid systemic disease | 515 (42.6%) | 2592 (53.8%) | 3107 (51.6%) | |
| Severe systemic disease | 679 (56.1%) | 2126 (44.1%) | 2805 (46.6%) | |
| Incapacitating disease | 8 (0.7%) | 32 (0.7%) | 40 (0.7%) | |
| Preoperative HICP status | 1005 (75.9%) | 1997 (37.6%) | 3002 (45.3%) | 0 |
| Preoperative anxiety diagnosis | 271 (26.4%) | 690 (17.3%) | 961 (19.1%) | 1610 |
| Preoperative depression diagnosis | 276 (26.9%) | 690 (17.3%) | 966 (19.2%) | 1610 |
| BMI, kg/m2 | 32.0 (6.4) | 30.5 (5.8) | 30.8 (5.9) | 725 |
| Joint | 0 | |||
| Hip | 487 (36.8%) | 1584 (29.9%) | 2071 (31.2%) | |
| Knee | 583 (44.0%) | 2508 (47.3%) | 3091 (46.6%) | |
| Shoulder | 254 (19.2%) | 1215 (22.9%) | 1469 (22.2%) | |
| Surgery type | 0 | |||
| Primary | 1120 (84.6%) | 4915 (92.6%) | 6035 (91.0%) | |
| Revision | 204 (15.4%) | 392 (7.4%) | 596 (9.0%) | |
| Discharge location | 0 | |||
| Home/self-care | 1064 (80.4%) | 4695 (88.5%) | 5759 (86.8%) | |
| Home health service | 141 (10.6%) | 340 (6.4%) | 481 (7.3%) | |
| Acute/rehabilitation/skilled nursing facility | 119 (9.0%) | 272 (5.1%) | 391 (5.9%) | |
| Hospital length of stay after surgery, d | 2.6 (2.6) | 1.9 (1.7) | 2.0 (2.0) | 4 |
| Preoperative alcohol consumption | 835 | |||
| Never/no | 344 (31.2%) | 1081 (23.1%) | 1425 (24.5%) | |
| Not currently | 363 (32.9%) | 1174 (25.0%) | 1537 (26.5%) | |
| Yes | 396 (35.9%) | 2438 (52.0%) | 2834 (48.9%) | |
| Preoperative tobacco use | 808 | |||
| Never/passive | 604 (54.4%) | 2772 (58.8%) | 3376 (58.0%) | |
| Quit | 432 (38.9%) | 1774 (37.6%) | 2206 (37.9%) | |
| Yes | 74 (6.7%) | 167 (3.5%) | 241 (4.1%) | |
| Preoperative opioid exposure | 565 (42.7%) | 1903 (35.9%) | 2468 (37.2%) | 0 |
| Postoperative opioid exposure | 858 (64.8%) | 2246 (42.3%) | 3104 (46.8%) | 0 |
Data presented as mean (standard deviation) or count (percentage).
ADI, area deprivation index; ASA, American Society of Anesthesiologists Classification; BMI, body mass index; HICP, high-impact chronic pain.
Table 2.
Pre- and postoperative patient-reported outcome measures (N = 6631).
| Measures | Pre-operative score | Post-operative score |
|---|---|---|
| PROMIS physical function | N = 6618 36.9 (6.9) | N = 6609 40.6 (8.0) |
| PROMIS pain interference | 64.1 (6.3) | 57.9 (8.4) |
| NPRS pain intensity | N = 4491 4.2 (3.0) | N = 3883 2.3 (2.7)* |
Data presented as mean (standard deviation).
Median (Q1-Q3) pain intensity was 1.0 (0.0–4.0), indicating postoperative data were skewed right.
PROMIS, patient-reported outcomes measurement information systems.
3.1. Pre and postoperative high-impact chronic pain status (aim 1)
For collective pre- and postoperative HICP, 45.3% (3002/6631) of patients reported preoperative HICP and 20.0% (1324/6631) were classified as having postoperative HICP. A small percentage of patients who did not report HICP preoperatively (4.8%, 319/6631) transitioned to HICP postoperatively, whereas 30.1% (1997/6631) of those who reported HICP preoperatively transitioned away from it postoperatively. Those who did not transition at all included 15.2% (1005/6631) who had HICP at both pre- and postoperative time points, and 49.9% (3310/6631) who did not report HICP at either pre- or postoperative assessments (Fig. 2).
Figure 2.
Sankey diagram depicting pre- and postoperative high-impact chronic pain transitions.
3.2. Factors associated with postoperative high-impact chronic pain (aim 2)
We evaluated candidate predictors for postoperative HICP in univariable and multivariable logistic regression models. Most candidate predictors were associated with postoperative HICP, except for sex (P = 0.13) and ethnicity (P = 0.82). Some variables showed increased odds in specific categories. For national ADI, the third (OR = 1.40, 95% 1.15–1.70) and fourth (OR = 1.92, 95% CI: 1.55–2.38) quartiles were at increased odds of postoperative HICP status, but not the second quartile, compared with the first quartile. Similarly, the perioperative factor of ASA score showed that only patients with an ASA status of severe systemic disease (OR = 2.64, 95% CI: 1.26–5.53) or incapacitating systemic disease (OR = 2.78, 95% CI: 1.07–7.24) were at increased odds of postoperative HICP compared with those with a healthy ASA status (Table 3).
Table 3.
Factors associated with postoperative HICP.
| Univariable | Multivariable | |||
|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | |
| Cohort demographics | ||||
| Age (linear spline at 72 y) | ||||
| 18 < age ≤72 | 0.97 (0.97–0.98) | <0.001 | 0.98 (0.98–0.99) | <0.001 |
| Age >72 | 1.04 (1.01–1.07) | 0.001 | 1.03 (0.99–1.06) | 0.10 |
| Sex | ||||
| Male | REF | REF | ||
| Female | 1.10 (0.97–1.24) | 0.13 | 0.83 (0.72–0.96) | 0.012 |
| Race | ||||
| White/Caucasian | REF | REF | ||
| Black/African American | 1.72 (1.49–1.99) | <0.001 | 1.16 (0.98–1.38) | 0.076 |
| Other | 1.48 (1.13–1.95) | 0.005 | 1.24 (0.91–1.68) | 0.17 |
| Ethnicity | ||||
| Not Hispanic/Latino | REF | |||
| Hispanic/Latino or other | 0.97 (0.72–1.30) | 0.82 | ||
| National ADI* | ||||
| First quartile | REF | REF | ||
| Second quartile | 1.07 (0.87–1.33) | 0.51 | 0.84 (0.67–1.07) | 0.16 |
| Third quartile | 1.40 (1.15–1.70) | <0.001 | 0.90 (0.72–1.13) | 0.38 |
| Fourth quartile | 1.92 (1.55–2.38) | <0.001 | 0.99 (0.77–1.27) | 0.93 |
| Preoperative risk factors | ||||
| Preoperative HICP status | ||||
| No | REF | REF | ||
| Yes | 5.22 (4.55–5.99) | <0.001 | 2.49 (2.10–2.94) | <0.001 |
| Pain intensity | 1.18 (1.15–1.21) | <0.001 | 1.07 (1.04–1.10) | <0.001 |
| PROMIS physical function score | 0.89 (0.88–0.90) | <0.001 | 0.95 (0.93–0.96) | <0.001 |
| BMI (linear spline at 26 kg/m2) | ||||
| 18 < BMI ≤26 | 0.98 (0.93–1.02) | 0.31 | 0.98 (0.93–1.03) | 0.43 |
| BMI >26 | 1.08 (1.02–1.14) | 0.006 | 1.03 (0.97–1.10) | 0.31 |
| ASA score | ||||
| Healthy | REF | |||
| Mild systemic disease | 1.66 (0.79–3.46) | 0.18 | 0.89 (0.31–2.59) | 0.83 |
| Severe systemic disease | 2.64 (1.26–5.53) | 0.010 | 1.17 (0.53–2.58) | 0.69 |
| Incapacitating systemic disease | 2.78 (1.07–7.24) | 0.036 | 1.33 (0.60–2.94) | 0.49 |
| Elixhauser sum score | 1.18 (1.15–1.21) | <0.001 | 1.06 (1.03–1.10) | <0.001 |
| Preoperative depression diagnosis | ||||
| No | REF | REF | ||
| Yes | 1.67 (1.44–1.94) | <0.001 | 1.11 (0.92–1.33) | 0.29 |
| Preoperative anxiety diagnosis | ||||
| No | REF | REF | ||
| Yes | 1.63 (1.39–1.90) | <0.001 | 1.17 (0.97–1.42) | 0.10 |
| Alcohol use | ||||
| No (never/no/not currently) | REF | REF | ||
| Yes | 0.54 (0.47–0.62) | <0.001 | 0.80 (0.68–0.94) | 0.006 |
| Tobacco use | ||||
| No (never/passive/quit) | REF | REF | ||
| Yes | 1.83 (1.38–2.42) | <0.001 | 1.25 (0.91–1.72) | 0.17 |
| Perioperative factors | ||||
| Joint site | ||||
| Hip | REF | REF | ||
| Knee | 0.76 (0.66–0.87) | <0.001 | 0.94 (0.80–1.10) | 0.43 |
| Shoulder | 0.68 (0.57–0.81) | <0.001 | 0.95 (0.78–1.16) | 0.62 |
| Surgery type | ||||
| Primary | REF | REF | ||
| Revision | 2.28 (1.91–2.74) | <0.001 | 1.78 (1.43–2.20) | <0.001 |
| Discharge location | ||||
| Home/self-care | REF | REF | ||
| Home health | 1.83 (1.49–2.25) | <0.001 | 0.92 (0.58–1.48) | 0.74 |
| Rehab/SNF/LTAC | 1.93 (1.54–2.42) | <0.001 | 0.84 (0.60–1.16) | 0.29 |
| Hospital length of stay after surgery (per day increase) | 1.16 (1.12–1.19) | <0.001 | 1.02 (0.98–1.06) | 0.36 |
| Opioid exposure | ||||
| Preoperative | ||||
| No | REF | REF | ||
| Yes | 1.33 (1.18–1.50) | <0.001 | 1.03 (0.90–1.18) | 0.67 |
| Postoperative | ||||
| No | REF | REF | ||
| Yes | 2.51 (2.21–2.84) | <0.001 | 1.83 (1.59–2.11) | <0.001 |
ADI, area deprivation index; HICP, high-impact chronic pain; SNF, Skilled Nursing Facility; LTAC. Long term care facility.
Next, we performed a multivariable logistic regression to identify the factors associated with postoperative HICP 6 months after arthroplasty. Key findings were that the strongest predictor for postoperative HICP was preoperative HICP Status (OR = 2.49, 95% CI: 2.10–2.94), followed by postoperative opioid exposure (OR = 1.83, 95% CI: 1.59–2.11). Additionally, preoperative pain intensity (OR = 1.07, 95% CI: 1.04–1.10) and Elixhauser summary score (OR = 1.06, 95% CI: 1.03–1.10) were associated with postoperative HICP. Age from 18 to 72 years (OR = 0.98 per year increase, 95% CI: 0.98–0.99), preoperative alcohol use (OR = 0.80, 95% CI: 0.68–0.94), and preoperative physical function (OR = 0.95 per unit increase, 95% CI: 0.93–0.96) were found to be protective of postoperative HICP. Race, ADI, anxiety, tobacco use, BMI, ASA status, depression, anxiety, preoperative opioid exposure, TJA site, discharge location, and hospital LOS were not statistically significant predictors in the final model, but revision surgery was associated with increased likelihood of HICP (OR = 1.78, 95% CI: 1.43–2.20) (Table 3). Our model correctly classified 81.4% of patients with HICP and demonstrated good discrimination in predicting postoperative HICP status, with an AUC of 0.78 (95% CI: 0.76–0.78) (Fig. 3).
Figure 3.
Performance of logistic regression model in predicting postoperative HICP using a predicted probability threshold = 50%.
4. Discussion
This study described transitions of HICP after TJA and identified the strongest risk factors contributing to HICP postoperatively, which were preoperative HICP status, postoperative opioid exposure, higher pain intensity, and lower physical function. Specifically, we found that nearly half of the patients (45.3%) reported preoperative HICP, and 20% continued to experience HICP following surgery, as illustrated in Figure 2. Approximately 15% of these patients had experienced persistent HICP preoperatively, and a lower proportion (5%) experienced new-onset HICP. Total joint arthroplasty is an effective intervention for OA and is indicated when patients meet the criteria for clinical severity24,42; however, there remains a sizable portion of patients who experience persistent pain.52 The findings from this cohort provide guidance on which factors were associated with an increase in postoperative HICP status, an outcome of high clinical importance that has not been widely reported in TJA populations.
The strongest preoperative predictor for postoperative HICP status was preoperative HICP status. The odds of postoperative HICP decreased from 5.22 in the univariable model to 2.49 in the multivariable model, suggesting that comorbidity burden, physical function, and pain intensity partially mediate this relationship. Even after controlling for preoperative HICP status, lower physical function and higher pain intensity remained strong independent predictors, increasing the odds of postoperative HICP by 5% for every 1-point reduction in physical function and by 7% for every 1-point increase in pain intensity. Postoperative opioid use was associated with an 83% increased risk for HICP after TJA (P < 0.001). However, preoperative exposure was not a significant predictor and may reflect its routine use in this population before surgery to manage symptoms. There is no indication from our data that continued opioid exposure postoperatively was inappropriate; instead, postoperative pain may have been severe enough to manage via opioid therapy. An alternative possibility would be the potential for opioid-induced hyperalgesia, where prolonged opioid exposure paradoxically heightens pain sensitivity.38 This could be compounded by the interplay of psychological factors leading to continued opioid exposure and persistent pain.30
We identified several similarities to previous literature but also noted some differences. For example, age, female sex, non-White race, Hispanic ethnicity, and socioeconomic disparities are established risk factors for HICP.8,14,29,31,34 In our study, each of these factors was an independent predictor in univariable analysis, except for sex and ethnicity. Interestingly, in the multivariable models, only younger age and sex were independent predictors; however, Black/African American race was associated with increased risk compared with White/Caucasian race, although it did not meet traditional thresholds for statistical significance (OR = 1.16, P = 0.076). This finding suggests that potential disparities related to access to care, implicit bias, or differences in pain communication may attenuate this relationship. Area deprivation index not being a strong contributor to these models could indicate that neighborhood-level socioeconomic disadvantage, when considered alongside other factors, may not play a crucial role in transitions in HICP. Preoperative measures of health status, such as higher BMI, increased comorbidity, lifestyle factors (ie, alcohol and tobacco use), and patient-reported preoperative pain (higher) and function (lower) have also been noted as risk factors and were found to be predictors in univariable analyses.14,45 However, in our adjusted models, BMI and ASA classification did not contribute to predicting postoperative HICP. This finding suggests that comorbidity indices, such as the Elixhauser index, may offer greater predictive utility and may better capture systemic health influences on HICP. Interestingly, there was a lack of association of preoperative diagnoses of anxiety and depression with postoperative HICP, where this relationship has been well established in the chronic pain literature. Other factors that have been associated with persistent pain include pain catastrophizing. Still, symptoms of anxiety and depression have the highest strength of evidence compared with other psychological risk factors in 1 systematic review.1,16,27,47 In this cohort, these factors did not contribute to the multivariable model. It is essential to note that several relevant psychological factors were not measurable in this study because they are not routinely collected in the EHR, including maladaptive pain-coping strategies, somatization of pain, and pain catastrophizing.46 Although many predictors increased the risk for postoperative HICP, we found that preoperative report of alcohol use modestly reduced the risk (20%) of developing postoperative HICP. This finding is consistent with prior studies; however, our dichotomous use of this variable does not elucidate the heterogeneous nature of alcohol consumption, where moderate alcohol use is linked to better outcomes, whereas excessive use may worsen outcomes.25,41 Therefore, this should be interpreted within this context and not as a causal relationship.25 Finally, other perioperative factors, including TJA site, discharge location, and hospital LOS, were not associated with postoperative HICP status. Still, revision surgeries confer an 80% higher risk of HICP compared with primary surgeries; a finding that was not consistent with our prior cross-sectional work but is supported in the literature across other longitudinal TJA cohorts investigating pain outcomes.33,43
This study has several study design elements that should be considered when interpreting its strengths and limitations. First, this study's cohort was large (n = 6631) and the first to examine HICP longitudinally in TJA. However, our cohort was limited to a single academic medical center, potentially limiting generalizability. Additionally, including both primary and revision TJA procedures enhances the clinical relevance of the results, as it captures a broader range of patients than primary TJA alone. Next, this was an observational study; therefore, we cannot speak to the effectiveness of TJA as an intervention for HICP. Relatedly, the primary outcome measure in this study was derived from clinical data, introducing variability in the timing of collection of the six-month follow-up data. This reflects the strengths and challenges of real-world data: it enhances generalizability and mirrors actual clinical practice but reduces temporal consistency and may lead to residual misclassification. Despite this limitation, our model classified 4 of 5 patients correctly and may be particularly valuable for identifying those at low risk. Rather than serving as a strict high-/low-risk dichotomy, this model can guide clinical judgment and resource allocation. For example, patients with preoperative HICP, high pain intensity, and low physical function may benefit most from multidisciplinary prehabilitation or targeted perioperative pain-management interventions.
In conclusion, although TJA substantially reduces HICP prevalence, 1 in 5 patients continues to experience significant postoperative pain, most of whom reported preoperative HICP. The findings of this study and the existing literature underscore the need to identify at-risk individuals and develop targeted, patient-centered perioperative interventions. Therefore, treatment approaches for these groups may need to differ, and pain management strategies should be tailored to be patient centered and evidence informed. Integrating treatment approaches such as medical optimization, prehabilitation, and cognitive-behavioral therapy may help prevent transitions to HICP, especially when coordinated by a multidisciplinary team.6,21,22,44
Disclosures
The authors have no conflict of interest to declare.
Supplemental digital content
Supplemental digital content associated with this article can be found online at http://links.lww.com/PR9/A385.
Supplementary Material
Acknowledgements
Funding: SG work on this manuscript was supported by a NIH/NIAMS grant (AR081796).
Author Contributions: MH, MR, SG, CG and MB were responsible for the conception and design of the study; MH, CG, SG were involved in the processing and statistical analysis of data; MH, MR, CG, MB and SG were involved in the drafting of the manuscript; and all authors contributed to the interpretation of the data for the work and revising it critically for important intellectual content. All the authors finally approved the manuscript. MH takes responsibility for the integrity of the work as a whole. All authors have read and agreed to the published version of the manuscript.
Data availability statement: All data will be made available to other investigators upon request to the corresponding author.
Institutional Board Approval: The Duke University Institutional Review Board approved this study (Pro00091740).
Footnotes
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Web site (www.painrpts.com).
Contributor Information
Michelle M. Ramirez, Email: michelle.ramirez@duke.edu.
Cynthia L. Green, Email: cindy.green@duke.edu.
Michael P. Bolognesi, Email: michael.bolognesi@duke.edu.
Steven Z. George, Email: steven.george@duke.edu.
References
- [1].Aaron RV, Ravyts SG, Carnahan ND, Bhattiprolu K, Harte N, McCaulley CC, Vitalicia L, Rogers AB, Wegener ST, Dudeney J. Prevalence of depression and anxiety among adults with chronic pain: a systematic review and meta-analysis: a systematic review and meta-analysis. JAMA Netw Open 2025;8:e250268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Amtmann D, Cook KF, Jensen MP, Chen W-H, Choi S, Revicki D, Cella D, Rothrock N, Keefe F, Callahan L, Lai J-S. Development of a PROMIS item bank to measure pain interference. PAIN 2010;150:173–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Blakeney WG, Kuster M. Advances in hip replacement surgery. J Clin Med 2023;12:3439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Calcaterra SL, Scarbro S, Hull ML, Forber AD, Binswanger IA, Colborn KL. Prediction of future chronic opioid use among hospitalized patients. J Gen Intern Med 2018;33:898–905. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Cheng H-Y, Beswick AD, Bertram W, Siddiqui MA, Gooberman-Hill R, Whitehouse MR, Wylde V. What proportion of people have long-term pain after total hip or knee replacement? An update of a systematic review and meta-analysis. BMJ Open 2025;15:e088975. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Coronado RA, Patel AM, McKernan LC, Wegener ST, Archer KR. Preoperative and postoperative psychologically informed physical therapy: a systematic review of randomized trials among patients with degenerative spine, hip, and knee conditions. J Appl Biobehav Res 2019;24:e12159. [Google Scholar]
- [7].Cunningham DJ, George SZ, Lewis BD. The impact of state level public policy, prescriber education, and patient factors on opioid prescribing in elective orthopedic surgery: findings from a tertiary, academic setting. Mayo Clin Proc Innov Qual Outcomes 2021;5:23–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Dahlhamer J, Lucas J, Zelaya C, Nahin R, Mackey S, DeBar L, Kerns R, Von Korff M, Porter L, Helmick C. Prevalence of chronic pain and high-impact chronic pain among adults - united States, 2016. MMWR Morb Mortal Wkly Rep 2018;67:1001–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Department of Health and Human Services. National pain strategy: a comprehensive population health-level strategy for pain. Washington, DC: National Pain Strategy, 2016. [Google Scholar]
- [10].Deyo RA, Dworkin SF, Amtmann D, Andersson G, Borenstein D, Carragee E, Carrino J, Chou R, Cook K, DeLitto A, Goertz C, Khalsa P, Loeser J, Mackey S, Panagis J, Rainville J, Tosteson T, Turk D, Von Korff M, Weiner DK. Report of the NIH task force on research standards for chronic low back pain. J Pain 2015;8:16–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Dripps RD, Lamont A, Eckenhoff JE. The role of anesthesia in surgical mortality. JAMA 1961;178:261–6. [DOI] [PubMed] [Google Scholar]
- [12].Duca LM, Helmick CG, Barbour KE, Nahin RL, Von Korff M, Murphy LB, Theis K, Guglielmo D, Dahlhamer J, Porter L, Falasinnu T, Mackey S. A review of potential national chronic pain surveillance systems in the United States. J Pain 2022;23:1492–509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Ethgen O, Bruyère O, Richy F, Dardennes C, Reginster J-Y. Health-related quality of life in total hip and total knee arthroplasty. A qualitative and systematic review of the literature. J Bone Joint Surg Am 2004;86:963–74. [DOI] [PubMed] [Google Scholar]
- [14].Fernández-de-Las-Peñas C, Florencio LL, de-la-Llave-Rincón AI, Ortega-Santiago R, Cigarán-Méndez M, Fuensalida-Novo S, Plaza-Manzano G, Arendt-Nielsen L, Valera-Calero JA, Navarro-Santana MJ. Prognostic factors for postoperative chronic pain after knee or hip replacement in patients with knee or hip osteoarthritis: an umbrella review. J Clin Med 2023;12:10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Fitz-Henry J. The ASA classification and peri-operative risk. Ann R Coll Surg Engl 2011;93:185–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Gatchel RJ, Peng YB, Peters ML, Fuchs PN, Turk DC. The biopsychosocial approach to chronic pain: scientific advances and future directions. Psychol Bull 2007;133:581–624. [DOI] [PubMed] [Google Scholar]
- [17].George SZ, Bolognesi MP, Bhavsar NA, Penrose CT, Horn ME. Chronic pain prevalence and factors associated with high impact chronic pain following total joint arthroplasty: an observational study. J Pain 2022;23:450–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].George SZ, Rubenstein D, Bolognesi MP, Horn ME. Can patient-reported outcome measurement information system measures estimate high impact chronic pain after total joint arthroplasty? J Arthroplasty 2023;38:S47–S51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Gunaratne R, Pratt DN, Banda J, Fick DP, Khan RJK, Robertson BW. Patient dissatisfaction following total knee arthroplasty: a systematic review of the literature. J Arthroplasty 2017;32:3854–60. [DOI] [PubMed] [Google Scholar]
- [20].Horn ME, George SZ, Goode AP, Reinke EK, Scott LG, Bolognesi MP. Can patient-reported outcome measurement information system measures differentiate patients who will undergo hip and knee total joint arthroplasty: a retrospective case-control study. J Arthroplasty 2022;37:S56–S62. [DOI] [PubMed] [Google Scholar]
- [21].Hunt MA, Keefe FJ, Bryant C, Metcalf BR, Ahamed Y, Nicholas MK, Bennell KL. A physiotherapist-delivered, combined exercise and pain coping skills training intervention for individuals with knee osteoarthritis: a pilot study. Knee 2013;20:106–12. [DOI] [PubMed] [Google Scholar]
- [22].Hurley MV, Walsh NE, Mitchell HL, Pimm TJ, Patel A, Williamson E, Jones RH, Dieppe PA, Reeves BC. Clinical effectiveness of a rehabilitation program integrating exercise, self-management, and active coping strategies for chronic knee pain: a cluster randomized trial. Arthritis Rheum 2007;57:1211–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Interagency Pain Research Coordinating Committee (IPRCC). National institute of neurological disorders and stroke. 2025. Available: https://www.iprcc.nih.gov/national-pain-strategy-overview/national-pain-strategy-full-report. Accessed July 10, 2025. [Google Scholar]
- [24].Kahlenberg CA, Nwachukwu BU, McLawhorn AS, Cross MB, Cornell CN, Padgett DE. Patient satisfaction after total knee replacement: a systematic review. HSS J 2018;14:192–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Karimi R, Mallah N, Nedjat S, Beasley MJ, Takkouche B. Association between alcohol consumption and chronic pain: a systematic review and meta-analysis. Br J Anaesth 2022;129:355–65. [DOI] [PubMed] [Google Scholar]
- [26].Kessler BA, Burrus B, Somashekar G, Wurzelmann SP, Bhowmick D. Limitations on postoperative opioid prescriptions and effects on health care resource use following elective anterior cervical discectomy and fusion. World Neurosurg 2021;146:e501–e508. [DOI] [PubMed] [Google Scholar]
- [27].Korff V, Ormel M, Keefe J, Dworkin FJ. Grading the severity of chronic pain. PAIN 1992;50:133–49. [DOI] [PubMed] [Google Scholar]
- [28].Kroenke K, Krebs EE, Turk D, Von Korff M, Bair MJ, Allen KD, Sandbrink F, Cheville AL, DeBar L, Lorenz KA, Kerns RD. Core outcome measures for chronic musculoskeletal pain research: recommendations from a veterans health administration work group. Pain Med 2019;20:1500–8. [DOI] [PubMed] [Google Scholar]
- [29].Luo D, Fan Z, Yin W. Chronic post-surgical pain after total knee arthroplasty: a narrative review. Perioper Med (Lond) 2024;13:108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Martinez-Calderon J, Flores-Cortes M, Morales-Asencio JM, Luque-Suarez A. Pain catastrophizing, opioid misuse, opioid use, and opioid dose in people with chronic musculoskeletal pain: a systematic review. J Pain 2021;22:879–91. [DOI] [PubMed] [Google Scholar]
- [31].Nahin RL, Feinberg T, Kapos FP, Terman GW. Estimated rates of incident and persistent chronic pain among US adults, 2019-2020. JAMA Netw Open 2023;6:e2313563. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Patil VV, Sancheti PK, Patil K, Gugale S, Shyam A. Functional outcome of mechanical alignment in total knee arthroplasty surgery: a short-term cohort study at an Indian tertiary care hospital. Indian J Orthop 2024;58:11–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Petersen KK, Simonsen O, Laursen MB, Nielsen TA, Rasmussen S, Arendt-Nielsen L. Chronic postoperative pain after primary and revision total knee arthroplasty. Clin J Pain 2015;31:1–6. [DOI] [PubMed] [Google Scholar]
- [34].Pitcher MH, Von Korff M, Bushnell MC, Porter L. Prevalence and profile of high-impact chronic pain in the United States. J Pain 2019;20:146–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].Rashiq S, Dick BD. Post-surgical pain syndromes: a review for the non-pain specialist. Can J Anaesth 2014;61:123–30. [DOI] [PubMed] [Google Scholar]
- [36].Ravindran D. Chronic postsurgical pain: prevention and management. J Pain Palliat Care Pharmacother 2014;28:51–3. [DOI] [PubMed] [Google Scholar]
- [37].Ryan Powell W, Sheehy AM, Kind AJH. The Area Deprivation Index Is The Most Scientifically Validated Social Exposome Tool Available For Policies Advancing Health Equity. Health Aff Forefront July 20, 2023 [epub ahead of print].DOI: 10.1377/forefront.20230714.676093 [DOI] [Google Scholar]
- [38].Sampaio-Cunha TJ, Martins I. Knowing the enemy is halfway towards victory: a scoping review on opioid-induced hyperalgesia. J Clin Med 2022;11:6161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [39].Sayah SM, Karunaratne S, Beckenkamp PR, Horsley M, Hancock MJ, Hunter DJ, Herbert RD, de Campos TF, Steffens D. Clinical course of pain and function following total knee arthroplasty: a systematic review and meta-regression. J Arthroplasty 2021;36:3993–4002.e37. [DOI] [PubMed] [Google Scholar]
- [40].Schug SA, Lavand’homme P, Barke A, Korwisi B, Rief W, Treede R-D, IASP taskforce for the classification of chronic pain. The IASP classification of chronic pain for ICD-11: chronic postsurgical or posttraumatic pain. PAIN 2019;160:45–52. [DOI] [PubMed] [Google Scholar]
- [41].Scott JR, Hassett AL, Schrepf AD, Brummett CM, Harris RE, Clauw DJ, Harte SE. Moderate alcohol consumption is associated with reduced pain and fibromyalgia symptoms in chronic pain patients. Pain Med 2018;19:2515–27. [DOI] [PubMed] [Google Scholar]
- [42].Shichman I, Roof M, Askew N, Nherera L, Rozell JC, Seyler TM, Schwarzkopf R. Projections and epidemiology of primary hip and knee arthroplasty in medicare patients to 2040-2060. JB JS Open Access 2023;8:e22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [43].Skou ST, Graven-Nielsen T, Rasmussen S, Simonsen OH, Laursen MB, Arendt-Nielsen L. Widespread sensitization in patients with chronic pain after revision total knee arthroplasty. PAIN 2013;154:1588–94. [DOI] [PubMed] [Google Scholar]
- [44].Sorel JC, Overvliet GM, Gademan MGJ, den Haan C, Honig A, Poolman RW. The influence of perioperative interventions targeting psychological distress on clinical outcome after total knee arthroplasty. Rheumatol Int 2020;40:1961–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [45].Sydora BC, Whelan LJ, Abelseth B, Brar G, Idris S, Zhao R, Leonard AJ, Rosenbloom BN, Clarke H, Katz J, Beesoon S, Rasic N. Identification of presurgical risk factors for the development of chronic postsurgical pain in adults: a comprehensive umbrella review. J Pain Res 2024;17:2511–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [46].Tanguay-Sabourin C, Fillingim M, Guglietti GV, Zare A, Parisien M, Norman J, Sweatman H, Da-Ano R, Heikkala E, Perez J, Karppinen J, Villeneuve S, Thompson SJ, Martel MO, Roy M, Diatchenko L, Vachon-Presseau E; PREVENT-AD Research Group. A prognostic risk score for development and spread of chronic pain. Nat Med 2023;29:1821–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [47].Theunissen M, Peters ML, Bruce J, Gramke H-F, Marcus MA. Preoperative anxiety and catastrophizing: a systematic review and meta-analysis of the association with chronic postsurgical pain. Clin J Pain 2012;28:819–41. [DOI] [PubMed] [Google Scholar]
- [48].Vervullens S, Meert L, Smeets RJEM, van der Nest G, Verbrugghe J, Verdonk P, Rahusen FTG, Meeus M. A biopsychosocial approach to phenotyping people with knee osteoarthritis awaiting total knee arthroplasty: a secondary cohort analysis. Ann Phys Rehabil Med 2024;67:101895. [DOI] [PubMed] [Google Scholar]
- [49].Vervullens S, Meert L, Smeets RJEM, Verbrugghe J, Verdonk P, Meeus M. Does pain intensity after total knee arthroplasty depend on somatosensory functioning in knee osteoarthritis patients? A prospective cohort study. Clin Rheumatol 2024;43:2047–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [50].van Walraven C, Austin PC, Jennings A, Quan H, Forster AJ. A modification of the elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care 2009;47:626–33. [DOI] [PubMed] [Google Scholar]
- [51].Wilson R, Pryymachenko Y, Audas R, Abbott JH. Long-term opioid medication use before and after joint replacement surgery in New Zealand. N Z Med J 2019;132:33–47. [PubMed] [Google Scholar]
- [52].Wylde V, Beswick A, Bruce J, Blom A, Howells N, Gooberman-Hill R. Chronic pain after total knee arthroplasty. EFORT Open Rev 2018;3:461–70. [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
Supplemental digital content associated with this article can be found online at http://links.lww.com/PR9/A385.