Abstract
Hip, knee, and shoulder arthroplasty are among the most frequently performed orthopaedic procedures in the United States. High impact and bothersome chronic pain rates following total joint arthroplasty (TJA) are unknown; as are factors that predict these chronic pain outcomes. This retrospective observational study included individuals that had a TJA from January 2014 to January 2020 (n = 2,638). Pre-operative and clinical encounter information was extracted from the electronic health record and chronic pain state was determined by email survey. Predictor variables included TJA location, number of surgeries, comorbidities, tobacco use, BMI, and pre-operative pain intensity. Primary outcomes were high impact and bothersome chronic pain. Rates of high impact pain (95% CI) were comparable for knee (9.8%-13.3%), hip (8.3%-11.8%) and shoulder (7.6%-16.3%). Increased risk of high impact pain included non-white race, two or more comorbidities, age less than 65 years, pre-operative pain scores 5/10 or higher, knee arthroplasty, and post-operative survey completion 24 months or less. Rates of bothersome chronic pain (95% CI) were also comparable for knee (24.9%-29.9%) and hip (21.3%-26.3%) arthroplasty; but higher for shoulder (26.9%-39.6%). Increased risk of bothersome chronic pain included non-white race, shoulder arthroplasty, knee arthroplasty, current or past tobacco use, and being female.
Introduction
Knee, hip, and shoulder arthroplasty are among the most frequently performed orthopaedic surgeries in the United States.15 Total knee and hip arthroplasty have had notable success in improving physical function and quality of life.18 However, overall effectiveness is limited by the 15-30% of patients that report chronic pain after total joint arthroplasty (TJA).32 This issue is not unique to joint arthroplasty as across all surgical procedures chronic postoperative pain is experienced by 10-50% of individuals.12, 13The individual impact of chronic postoperative pain is characterized by decreased physical functioning and quality of life, psychological distress, continued health care utilization, and long-term opioid usage.2, 22, 23, 31 Chronic postoperative pain has become more widely recognized as evidenced by it being added as a stand-alone diagnostic category in the upcoming 11th revision of the International Classification of Diseases.25
The Federal Pain Research Strategy has emphasized updated definitions of chronic pain, including emphasizing the importance of identifying high impact or bothersome chronic pain.17 These are chronic pain states characterized by disruptions in an individual’s activities of daily living, social roles, and ability to work due to pain.17 Recent population based investigations of high impact chronic pain did not include estimates for postoperative pain.5 Thus, additional research is needed to understand how quality of life is impacted by chronic pain following TJA. The purpose of this study was to report prevalence rates of high impact and bothersome chronic pain in patients that received a TJA from an academic health center. Our first aim was to determine how high impact and bothersome chronic pain rates differed based on location of TJA (i.e. hip, knee, or shoulder). Our second aim was to determine pre-operative or surgical encounter factors associated with high impact or bothersome chronic pain after TJA. In conjunction with these aims we investigated self-reported non-opioid or opioid pain medication use. This study will inform future direction in clinical practice by providing estimates of chronic pain following TJA using definitions aligned with the Federal Pain Research Strategy and by identifying pre-operative and clinical factors associated with increased risk of high impact and bothersome chronic pain after TJA.
Methods
Study Overview
This was a retrospective, observational study of patients who underwent a hip, knee, or shoulder TJA or TJA at any Duke Health affiliated hospital or ambulatory surgery clinic and completed an email survey regarding current chronic pain state. All patients received their TJA from an orthopaedic surgeon from Duke University Department of Orthopaedic Surgery. This study was approved by the Duke University Institutional Review Board and this paper was reported following the STROBE statement.28
Participants
Eligible patients were identified through the electronic health record (EHR, Epic Systems) using a starting date of January 1, 2014 (i.e. the time in which there was widespread use of the EHR for this patient population) and an end date of January 31, 2020. The end date was selected to allow for at least 6-months from post-operative to first survey time (i.e. the minimum period for development of chronic pain). We identified 17,338 patients who underwent a TJA or TJA related procedures during this time period using CPT codes (Supplemental File 1). Patients were excluded if they had died prior to January 31, 2020 (n=515), opted out of being contacted for research (n=138), did not have an email address on file (n=1,897), or had an invalid email address (n=1,257). After exclusions a total of 13,531 patients were eligible to participate in the study (Figure 1).
Figure 1. Study Flow for Survey Contact and Completion.
Figure 1 shows the total number of patients receiving total joint arthroplasty, the different numbers excluded, and the number that responded to the chronic pain survey.
Survey
An email survey was designed to collect self-reported information on chronic pain state, pain interference, and self-reported pain medication use. The eligible cohort was divided into 7 survey groups based on surgical year. Survey groups were created to allow for controlled distribution of the survey, which facilitated monitoring or response rates and the opportunity to respond to questions in a timely manner. If patients underwent more than one surgery, the first surgery date was used for survey group assignment. For each survey group a random number generator was used to determine survey contact date. Random assignment for survey administration was used to avoid influence of ordering effects that could confound our pain outcomes. Patients were then sorted in ascending order by participant ID within each randomly determined group. The survey was administered in weekly waves beginning July 06, 2020 and ending on November 06, 2020.
Participants were contacted via email to provide informed consent to participate in the survey. The email indicated that the reason for this survey was to assess joint pain after receiving a TJA from Duke Health. If patients did not complete the questionnaire on the first contact, two email reminders were sent three days apart. Additionally and in order to increase response rates, the research coordinator contacted patients who consented to participate in the survey but had not initiated or completed the survey. In these cases the research coordinator made a phone call to encourage survey completion. All survey information was collected by a link provided in the email that provided access to the secured survey.
Measures
Covariate and Exposure Ascertainment
Information extracted from the EHR prior to surgery included age, sex, self-identified race, body mass index (BMI), tobacco history, and comorbidity count from the EHR. We also extracted surgical encounter information related to the TJA including pre-operative pain rating (0-10 scale) within 30 days prior to surgery, the date of surgery, and the number of TJA or TJA related surgeries performed within the data extraction period. These variables were selected based on having clinical relevance (e.g. pre-operative pain intensity, number of surgeries) and/or prior association with TJA outcomes (e.g. BMI, age, tobacco use, comorbidity count, and sex). The length of time from first surgery to survey completion was selected to account for the variability in post-operative time.
Chronic Pain Grade
The development of the Graded Chronic Pain Scale has been previously described.30 In this cohort chronic pain was characterized using the first two questions from the Graded Chronic Pain Scale Revised (Q1: In the past 3 months, how often did you have pain? and Q2: Over the past 3 months, how often did pain limit your life or work activities?)29 and the Patient Reported Outcomes Measurement System (PROMIS®) measure for pain interference.1 In addition to these two questions, we used a short form instrument for the Pain Interference domain to determine Graded Chronic Pain Scale category. PROMIS measures are scored on a T-score metric, with standard population values for mean scores (50) and standard deviation (10). This population mean score of 50 was used to differentiate between mild (< 50 PROMIS pain interference score) and bothersome (>/= 50 PROMIS pain interference score) chronic pain grades. Thus when combined with the first two question responses, a score of 40 would have resulted in a mild chronic pain category, while a score of 60 would have resulted in a bothersome chronic pain category.
Accordingly, chronic pain categories were derived for each participant based on the following:
Chronic Pain Absent (Q1 Response = Never or Some days)
Mild Chronic Pain (Q1 Response = Most or Every day, Q2 Response = Never or Some days and PROMIS Pain Interference score < 50)
Bothersome Chronic Pain (Q1 Response = Most or Every day, Q2 Response = Never or Some days and PROMIS Pain Interference score >/= 50)
High Impact Chronic Pain (Q1 Response = Most or Every day, Q2 Response = Most or Every day)
Medication Use
Participants indicated current pain medication use (i.e. at the time of email survey) by reporting type of medication used, dosage prescribed, and how frequently it was used.
Data Management
Survey data were collected and managed using REDCap electronic data capture tools hosted at Duke University. REDCap (Research Electronic Data Capture) is a secure, web-based software platform designed to support data capture for research studies, providing 1) an intuitive interface for validated data capture; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for data integration and interoperability with external sources.7, 8
Data Analysis
Data analyses were completed with SPSS (IBM SPSS Statistics for Windows, Version 25.1, Armonk, NY: IBM Corp.). For the first aim, rates of high impact and bothersome chronic pain were calculated by arthroplasty location and reported as point and interval estimates (95% CI). The interval estimates were used to determine if high impact and bothersome chronic pain rates differed by arthroplasty location. Pre-operative and surgical continuous data (e.g. age, BMI, pre-operative pain intensity, comorbidity, and time from surgery to survey) were then collapsed to dichotomous variables to facilitate clinical interpretation. The cut-offs used to dichotomize these variables were determined a priori (i.e. before the data were analyzed) and based on input from the surgical collaborators for determining clinical relevance. This results in cut-offs of a) 24 months for time from surgery to survey (expectation of 2 year follow up data for TJA as a research standard); b) 4/10 pre-operative pain intensity (corresponding with higher pain intensity); c) 65 years for age (insurance eligibility); d) BMI of 30 (beginning range for obesity); and d) Comorbidity count of 2 (beginning range for multi-morbidity). These data were then reported descriptively by TJA location in the same categories as in the logistic regression models and frequencies were compared by chi-square tests.
The data for self-reported use of pain medication was collapsed into dichotomous measures (Y or N) for data analyses due to the distribution of responses. That is, there was a high frequency of patients reporting no pain medication use, especially for opioids (95.8%). These frequencies were compared by arthroplasty location using chi-square tests.
Finally, logistic regression models were used to determine if preoperative and surgical encounter variables were associated with increased odds of: 1) high impact or bothersome chronic pain (primary analysis, 2 distinct models) and 2) opioid use (secondary analysis, 1 model). The rationale for investigating opioid was based on prior work in spinal pain which indicated high impact chronic pain grades were strongly associated with long term opioid use.9 We wanted to see if a similar association was present for TJA. There was no correction for multiple comparisons and 95% CI was used for all comparisons.
Results
During the study period, 2,638 individuals (15.2% of those receiving TJA from January 2014 through January 2020) provided informed consent and completed the email survey (Figure 1). Survey participants did not differ based on age when compared to all individuals receiving TJA in this period (64.4 (sd = 9.6) vs. 64.7 (sd = 11.1) years, p = 0.17). However, there were differences in survey participants for other variables. Male participants were more likely to respond than female participants (16.6% vs. 14.1%, respectively), individuals identifying as Caucasian/While were more likely to respond than those identifying as Non-White (17.9% vs. 5.6%, respectively), participants with hip and knee surgery were more likely to respond than participants with shoulder surgery (15.9% hip and 15.2% knee vs. 12.1% shoulder), and participants with 2 or more TJA related surgeries were more likely to respond than participants with 1 surgery (17.1% 2+ surgeries vs. 14.7% 1 surgery) (all p<0.001).
Descriptive statistics are reported in Table 1 and rates of chronic pain by TJA location in Figure 2. Rates of high impact pain (95% CI) were comparable for knee (9.8%-13.3%), hip (8.3%-11.8%) and shoulder (7.6%-16.3%) arthroplasty. Rates of bothersome chronic pain (95% CI) were also comparable for knee (24.9%-29.9%), hip (21.3%-26.3%) and shoulder (26.9%-39.6%). Rates of self-reported pain medication use are reported in Table 2. Non-opioid pain medication use for all TJA was highest in the high impact (58.9%) and bothersome (53.7%) chronic pain categories (p < 0.001). Accordingly, opioid use for all TJA was highest in the high impact (15.4%) and bothersome (6.3%) chronic pain categories (p < 0.001). Self-reported non-opioid and opioid pain medication use was also strongly associated with individual TJA locations (Table 2).
Table 1.
Descriptive Statistics by Total Joint Arthroplasty Location
| All Arthroplasty (n = 2638) |
Total Knee (n = 1265) |
Total Hip (n =1146) |
Total Shoulder (n = 227) |
p-value | |
|---|---|---|---|---|---|
| Sex (% Female) |
53.2% (1404/2638) |
54.3% (687/1265) |
53.7% (615/1146) |
44.9% (102/227) |
0.031 |
| Tobacco Use (% No History) |
56.1% (1299/2316) |
56.9% (625/1099) |
56.9% (577/1014) |
47.8% (n = 97/203) |
0.044 |
| Race (% White/Caucasian) |
92.2% (2432/2638) |
92.1% (1165/1265) |
91.6% (1050/1146) |
95.6% (217/227) |
0.124 |
| Number Surgeries (% 1 Surgery) |
77.2% (2037/2638) |
74.0% (936/1265) |
79.3% (909/1146) |
84.6% (n = 192/227) |
< 0.001 |
| Survey Time (% > 24 months) |
72.0% (1897/2635) |
72.4% (915/1263) |
72.6% (831/1145) |
66.5% (151/227) |
0.158 |
| Pre-Operative Pain (% ≤ 4/10) |
61.3% (1517/2473) |
65.5% (797/1216) |
56.7% (599/1056) |
60.2% (121/201) |
<0.001 |
| Age (%: ≥ 65 years) |
55.7% (1469/2638) |
59.3% (750/1265) |
48.6% (557/1146) |
71.3% (162/227) |
<0.001 |
| BMI (% < 30) |
54.9% (1362/2480) |
47.0% (559/1190) |
63.1% (675/1070) |
58.2% (128/220) |
< 0.001 |
| Comorbidity (% 0-1) |
86.3% (2277/2638) |
85.9% (1087/1265) |
88.0% (1009/1146) |
79.7% (181/227) |
0.003 |
• P-values for frequency differences across arthroplasty location (via chi-square)
• Variables reported dichotomously to match how entered into logistic regression models
Figure 2. Chronic Pain Rates by Total Joint Arthroplasty Location.
Figure 2 shows rates of no, mild, bothersome, and high impact chronic pain for all total joint arthroplasty (on the left) and for knee, hip, and shoulder arthroplasty individually (in order from left to right).
Table 2.
Pain Medication Use by Total Joint Arthroplasty and Chronic Pain Grade
| Self-Reported Non-Opioid Medication Use | ||||
|---|---|---|---|---|
| All Arthroplasty (n = 2638) |
Total Knee (n = 1265) |
Total Hip (n = 1146) |
Total Shoulder (n = 227) |
|
| Chronic Pain Absent | 38.7% (455/1175) |
41.8% (216/517) |
35.7% (202/566) |
40.2% (37/92) |
| Mild Chronic Pain | 48.0% (233/485) |
51.4% (132/257) |
41.2% (80/194) |
61.8% (21/34) |
| Bothersome Chronic Pain | 53.7% (372/693) |
57.8% (200/346) |
50.4% (137/272) |
46.7% (35/75) |
| High Impact Chronic Pain | 58.9% (168/285) |
62.1% (90/145) |
53.5% (61/114) |
65.4% (17/26) |
| p-value | < 0.001 | < 0.001 | < 0.001 | 0.046 |
| Self-Reported Opioid Medication Use | ||||
| All Arthroplasty (n = 2638) |
Total Knee (n = 1265) |
Total Hip (n = 1146) |
Total Shoulder (n = 227) |
|
| Chronic Pain Absent | 1.4% (17/1175) |
1.5% (8/517) |
1.2% (7/566) |
2.2% (2/92) |
| Mild Chronic Pain | 1.2% (6/485) |
1.2% (3/257) |
1.5% (3/194) |
0.0% (0/34) |
| Bothersome Chronic Pain | 6.3% (44/693) |
7.8% (27/346) |
5.1% (14/272) |
4.0% (3/75) |
| High Impact Chronic Pain | 15.4% (44/285) |
13.1% (19/145) |
14.9% (17/114) |
30.8% (8/26) |
| p-value | < 0.001 | < 0.001 | < 0.001 | < 0.001 |
• P-values for frequency differences in chronic pain grade within each arthroplasty location (via chi-square)
Preoperative and surgical encounter variables associated with high impact or bothersome chronic pain (reference group = no or mild chronic pain for both models) are reported in Table 3. Variables included in the logistic regression model were examined for extent of missingness (Supplemental File 2). The majority of variables, including both outcome measures, had 0% missing values. Missing values exceeded 5% for 3 variables extracted from the EHR: BMI (6.0%), preoperative pain intensity (6.3%), and history of tobacco use (12.2%). Means and standard deviations were similar to participants with missing data (Supplemental File 2); subsequent analyses included patients with complete data.
Table 3.
Pre-Operative Factors Associated with High Impact or Bothersome Chronic Pain after Total Joint Arthroplasty
| Variable | High Impact Chronic Pain (n = 216 cases, n = 1273 controls) |
Bothersome Chronic Pain (n = 520 cases, n = 1273 controls) |
|---|---|---|
| Knee arthroplasty (ref: Hip) |
1.48
(95% CI = 1.07-2.04) |
1.40
(95% CI = 1.12-1.76) |
| Shoulder arthroplasty (ref: Hip) |
1.56 (95% CI = 0.89-2.75) |
1.88
(95% CI = 1.29-2.74) |
| Female (ref: Male) |
1.09 (95% CI = 0.81-1.47) |
1.36
(95% CI = 1.10-1.68) |
| Tobacco current or past (Ref: No History) |
1.30 (95% CI = 0.96-1.76) |
1.25
(95% CI = 1.01-1.55) |
| Race (Ref: White/Caucasian) |
2.00
(95% CI = 1.22-3.28) |
2.13
(95% CI = 1.47-3.08) |
| More than one surgery (Ref: One surgery) |
0.76 (95% CI = 0.48-1.20) |
0.99 (95% CI = 0.73-1.34) |
| Time from surgery to survey (Ref: more than 24) |
1.40
(95% CI = 1.02-1.93) |
1.13 (95% CI = 0.90-1.42) |
| Pre-operative pain score (Ref: 4/10 or lower) |
1.53
(95% CI = 1.13-2.07) |
1.08 (95% CI = 0.94-1.46) |
| Age at first surgery (Ref: 65 or older) |
1.58
(95% CI = 1.15-2.16) |
1.17 (95% CI = 0.68-1.06) |
| BMI (Ref: less than 30.0) |
1.31 (95% CI = 0.97-1.77) |
1.11 (95% CI = 0.89-1.37) |
| Comorbidity (Ref: 0-1) |
1.80
(95% CI = 1.23-2.65) |
1.19 (95% CI = 0.88-1.61) |
• Completed cases approach was used after missing values analyses indicated patterns of missing at random
• Sample size for cases and controls correspond with total number of individuals providing data for the regression model. In both models controls were the same individuals
• Odds ratios are adjusted for all other variables
Variables associated with increased risk of high impact pain were non-white race (aOR = 2.00; 95% CI = 1.22-3.28), two or more comorbidities (aOR = 1.80; 95% CI = 1.23-2.65), age less than 65 years (aOR = 1.58; 95% CI = 1.15-2.16), preoperative pain scores 5/10 or higher (aOR = 1.53, 95% CI = 1.13-2.07), knee arthroplasty (aOR = 1.48, 95% CI = 1.07-2.04), and post-operative survey completion 24 months or less (aOR = 1.40, 95% CI = 1.02-1.93). Variables associated with increased risk of bothersome chronic pain were non-white race (aOR = 2.13, 95% CI = 1.47-3.08), shoulder arthroplasty (aOR = 1.88, 95% CI = 1.29-2.74), knee arthroplasty (aOR = 1.40, 95% CI = 1.12-1.76), being female (aOR = 1.36, 95% CI = 1.10-1.68), and current or past tobacco use (aOR = 1.25, 95% CI = 1.01-1.55). Having two or more comorbidities was the only pre-operative variable associated with increased risk (aOR = 2.17; 95% CI = 1.30-3.64) in the model for postoperative opioid pain medication use.
As a follow-up to our planned analyses individuals with 2+ surgeries were excluded and parallel analyses were completed. High impact and bothersome chronic pain rates did not appreciably change with these individuals removed. Similarly, logistic regression estimates did not appreciably change. Therefore, in this manuscript we presented findings from our original analysis plan that included those with 2+ surgeries in all analyses.
Discussion
This findings of this observational study suggest that the prevalence of high impact chronic pain is approximately 10% rate following TJA. Moreover, postoperative bothersome chronic pain occurred at approximately a 25% rate following TJA. These rates for high impact and bothersome chronic pain did not differ appreciably based on TJA location. In adjusted analyses, non-white race and knee arthroplasty were individually associated with increased risk of both high impact and bothersome chronic pain following TJA. Pre-operative pain, which is a well-established risk factor for postoperative chronic pain,24, 33 was only associated with high impact chronic pain. These novel findings provide updated estimates of postoperative chronic pain rates following TJA using standardized definitions aligned with the Federal Pain Research Strategy.17 Additionally, factors associated with increased risk for chronic pain were identified and these factors could be used to develop decision support tools that inform patients and surgeons on the risk of chronic pain following TJA.
Prior studies of high impact chronic pain have not focused on postoperative populations, but do provide comparative data. Rates of high impact chronic pain from the National Health Interview Survey ranged from 4.8% (2011)21 to 8.0% (2016).5 Our findings suggest that higher rates of high impact pain can be expected following TJA. This is not a surprising finding given that TJA is a procedure used for those with significant limitations due to moderate or severe osteoarthritis. A higher rate of high impact chronic pain (15.0%) was reported in a cohort of healthcare plan enrollees; however these same enrollees had a much lower rate (10.8%) of bothersome chronic pain.29 The combined chronic pain rates (high impact and bothersome) are much higher from this TJA cohort, giving additional support to the expectation that the postoperative chronic pain rates will be higher than rates from non-operative cohorts. A high priority clinical research question for future consideration is whether chronic pain rates decrease following TJA. This research question can only be addressed with a prospective design.
This study supports use of the Revised Chronic Pain Grading Scale to classify TJA patients postoperatively.29 High impact and bothersome chronic pain grades were strongly associated with self-reported use of non-opioid and opioid pain medications, similar to a spinal pain cohorts investigating opioid use and high impact chronic pain.9 Identification of high impact or bothersome chronic pain could be a valuable addition to TJA outcome assessment via registry or EHR. This classification may be helpful in supporting clinical decision for whether additional pain management or postoperative rehabilitation is indicated. An additional advantage of such classification is that patients with TJA in different locations can be compared on the same metric.29 Furthermore, use of this metric allows for comparison of chronic pain rates to patient and general populations. As such, we recommend incorporating these items into future TJA outcome batteries to better characterize patient reported pain outcomes.
These are distinct clinical populations so the preoperative and surgical encounter differences across the TJA locations were an expected finding. In the multivariable regression models several factors were identified that could refine individual risk of high impact (e.g. non-white race, knee arthroplasty, pre-operative pain intensity 5/10 or higher, age less than 65 years, and having more than 1 comorbidity) or bothersome (e.g. non-white race, knee or shoulder arthroplasty, being female, and a history of tobacco use) chronic pain. Having more than 1 comorbidity was the only individual factor associated with increased risk of opioid use following TJA. International TJA cohorts that used similar predictors provide context for these findings. Increased comorbidity counts appear to be a consistent predictor of poor TJA outcomes; as this finding was also reported in cohorts from the Netherlands26, 27 and Canada6. However, an exception to the role comorbidities play for TJA outcomes was also identified, as a different Canadian cohort reported that number of comorbidities did not limit the improvement for pain, function, and acceptable symptom state.14 Interestingly, in a cohort from Greece females reported lower quality of life following total knee arthroplasty19, converging with our finding that females were at increased risk of reporting bothersome chronic pain. Comparisons to cohorts outside the United States must be tempered somewhat by the use of different outcome measures (i.e. none we identified used high impact or bothersome chronic pain). Future research involving high impact or bothersome chronic pain as outcomes in international cohorts will allow for more direct comparisons of predictor variables.
Two factors were predictive of both high impact and bothersome chronic pain - knee arthroplasty and non-white race. The implication for knee arthroplasty is that decision support for surgical selection may have to include individual risk of long term pain outcomes, in addition to what is typically considered (e.g. hospital readmission, infection risk, and need for revision). Non-white race had the largest non-response rates for the survey and strongest association with both chronic pain outcomes; both of which were unexpected findings. Similar health related disparities have been identified in TJA for utilization and short term medical outcomes;16, 20, 34 as have racial biases in medical providers that lead to disparities in pain assessment and treatment.10 Additionally, black women have lower physical functioning following total knee arthroplasty when compared to white women.3 This finding is attributed to differences in lower pre-operative functioning3, rather than differential utilization of post-operative rehabilitation between black and white women.4 Our findings converge with these reports and add to this literature by being the first we are aware of that identified race related differences in chronic pain outcomes following TJA. The study of race in medical research is complex;11 and this study does not directly address these complexities with the way race was categorized (i.e. self-identification) or by being unable to address confounding from other relevant variables (e.g. socioeconomic status). However, this study does suggest that race related disparities in chronic pain outcomes following TJA should be considered an area of emphasis in future studies.
Surveying all patients within a single health system, using updated definitions of chronic pain, and reporting on 3 different TJA procedures are notable strengths of this study. There are also several limitations to consider. First, the cross-sectional nature of the survey did not allow us to identify predictors of high impact or bothersome chronic pain. Second, we were unable to include pre-operative pain duration in the regression models because this variable was not available in the EHR. Therefore, we were unable to determine if pre-operative pain duration was associated with post-operative chronic pain status. Related, there was high variability in the postoperative time for completing the survey. Specifically, survey time of less than 24 months postoperatively was associated with increased risk of high impact chronic pain, but not bothersome chronic pain. The clinical implications of the survey time findings are limited but do suggest that postoperative follow up time can potentially influence the prevalence rates of chronic pain following TJA. Third, the overall survey response rate was low and there were differences in demographic variables for those that responded. Two caveats to consider when interpreting the survey response rates are: 1) this was a conservative estimate as it was calculated from all individuals having TJA (i.e. not just those eligible to participate in the survey) and 2) it canvassed post-operative responses over almost 6-years of TJA. Fourth, we used the PROMIS Pain Interference to differentiate bothersome chronic pain from mild chronic pain, when using the Pain Enjoyment of Life and General Activity Pain Interference Scale is the preferred method. Therefore, we have no direct post-operative measure of pain intensity for this cohort which could be considered a limitation. Fifth, there were missing data when variables were extracted from the EHR, most likely due to incomplete entry. These missing data limited the number of participants included in the logistic regression models. Sixth, participants completing the survey knew they were being contacted because they had a TJA but we did not explicitly instruct them to limit their responses to pain related to the surgery. Therefore, pain from other joints or body areas could be contributing to the estimates reported in this analysis.
The study also generated directions for future research. These findings could be used to design a prospective study that determines whether the factors identified in this analysis provide accurate risk estimates for high impact or bothersome chronic pain outcomes following TJA. Another direction for future research would be to include modifiable factors (e.g. psychosocial distress, biological markers) in prediction models for future prospective studies, since all the factors included in this analysis were not readily modifiable. One example of a modifiable factor, pre-operative pain catastrophizing, was predictive of 24-month pain outcomes following total knee arthroplasty.6 Finally, we did not include interaction terms in our regression models (e.g. age by surgical location) and future research could determine whether there are interactions present among the variables we identified as predictors of these chronic pain outcomes.
Conclusion
These findings provide estimates of chronic pain prevalence following TJA using standardized definitions aligned with the Federal Pain Research Strategy.17 For all TJA locations the rates of high impact and bothersome chronic pain were approximately 10% and 25% respectively. Individual factors with increased risk of both high impact and bothersome chronic pain were non-white race and knee arthroplasty. In this cohort, higher preoperative comorbidity count was the only factor associated with increased risk of opioid use following TJA.
Supplementary Material
Highlights.
High impact chronic pain rates were similar following hip, knee, and shoulder arthroplasty
Shoulder arthroplasty had higher rates of bothersome chronic pain, compared to hip arthroplasty
Only non-white race and knee arthroplasty were associated with both chronic pain outcomes
Perspective:
In this cohort more than 1/3rd of individuals reported high impact or bothersome chronic pain following TJA. Non-white race and knee arthroplasty were the only two variables associated with both chronic pain outcomes.
Disclosures:
The Duke Department of Orthopaedic Surgery Total Joint Arthroplasty Learning Health Unit provided support for the statistical analysis. SG received used salary support from NIH/NIAMS (AR055899) while working on writing and data analysis for this manuscript. These funders had no direct oversite of this manuscript and there are no other sources of funding to report. Each author certifies that he or she has no commercial associations (eg, consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might pose a conflict of interest in connection with the submitted article. Outside of this work SG reports personal fees from Rehab Essentials, Inc, and personal fees from Med Risk, LLC and MB reports personal fees from TJO, personal fees from Zimmer, grants and personal fees from Biomet, other from EOA, other from AAHKS, grants from Depuy, grants from Exactech, NB reports grants from NIH.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Ethical Review Committee Statement: This study was determined to be exempt by the Duke University Institutional Review Board.
Contributor Information
Steven Z. George, Department of Orthopaedic Surgery and Duke Clinical Research Institute, Duke University; 200 Morris Street, Durham NC 27001.
Michael P. Bolognesi, Department of Orthopaedic Surgery, Division of Adult Reconstruction, Duke University, Durham NC); 311 Trent Drive Durham, NC 27710.
Nrupen A. Bhavsar, Department of General Internal Medicine, Duke University, 200 Morris Street, Durham NC 27001
Colin T. Penrose, Department of Orthopaedic Surgery, Division of Adult Reconstruction, Duke University, Durham NC); 311 Trent Drive Durham, NC 27710.
Maggie E. Horn, (Department of Orthopaedic Surgery, Division of Physical Therapy, Duke University, Durham NC); 311 Trent Drive Durham, NC 27710
References
- 1.Amtmann D, Cook KF, Jensen MP, Chen WH, Choi S, Revicki D, Cella D, Rothrock N, Keefe F, Callahan L, Lai JS. Development of a PROMIS item bank to measure pain interference. Pain 150:173–182, 2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Calcaterra SL, Scarbro S, Hull ML, Forber AD, Binswanger IA, Colborn KL. Prediction of Future Chronic Opioid Use Among Hospitalized Patients. Journal of General Internal Medicine 33:898–905, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Cavanaugh AM, Rauh MJ, Thompson CA, Alcaraz J, Mihalko WM, Bird CE, Corbie-Smith G, Rosal MC, Li W, Shadyab AH, Gilmer T, LaCroix AZ. Racial/Ethnic Disparities in Physical Function Before and After Total Knee Arthroplasty Among Women in the United States. JAMA Netw Open 3: e204937, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cavanaugh AM, Rauh MJ, Thompson CA, Alcaraz JE, Bird CE, Gilmer TP, LaCroix AZ. Rehabilitation After Total Knee Arthroplasty: Do Racial Disparities Exist? The Journal of Arthroplasty 35:683–689, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.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. Morbidity and Mortality Weekly Report 67:1001–1006, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Forsythe ME, Dunbar MJ, Hennigar AW, Sullivan MJ, Gross M. Prospective relation between catastrophizing and residual pain following knee arthroplasty: two-year follow-up. Pain Research & Management 13:335–341, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Harris PA, Taylor R, Minor BL, Elliot V, Fernandez M, O’Neal L, McLeod L, Delacqua G, Delacqua F, Kirby J, Duda SN, REDCap Consortium. The REDCap consortium: Building an international community of software platform partners. Journal of Biomedical Informatics 95:103208, 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. Journal of Biomedical Informatics 42:377–381, 2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Herman PM, Broten N, Lavelle TA, Sorbero ME, Coulter ID. Health Care Costs and Opioid Use Associated With High-impact Chronic Spinal Pain in the United States. Spine 44:1154–1161, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hoffman KM, Trawalter S, Axt JR, Oliver MN. Racial bias in pain assessment and treatment recommendations, and false beliefs about biological differences between blacks and whites. Proceedings of the National Academy of Sciences of the United States of America 113: 4296–4301, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ioannidis JPA, Powe NR, Yancy C. Recalibrating the Use of Race in Medical Research. JAMA 325:623–624, 2021 [DOI] [PubMed] [Google Scholar]
- 12.Johansen A, Romundstad L, Nielsen CS, Schirmer H, Stubhaug A. Persistent postsurgical pain in a general population: prevalence and predictors in the Tromsø study. Pain 153:1390–1396, 2012. [DOI] [PubMed] [Google Scholar]
- 13.Kehlet H, Jensen TS, Woolf CJ. Persistent postsurgical pain: risk factors and prevention. Lancet (London, England) 367:1618–1625, 2006 [DOI] [PubMed] [Google Scholar]
- 14.King LK, Waugh EJ, Jones CA, Bohm E, Dunbar M, Woodhouse L, Noseworthy T, Marshall DA, Hawker GA, BEST-Knee Study Team. Comorbidities do not limit improvement in pain and physical function after total knee arthroplasty in patients with knee osteoarthritis: the BEST-Knee prospective cohort study. BMJ Open. 11:e047061. 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Molina CS, Thakore RV, Blumer A, Obremskey WT, Sethi MK. Use of the National Surgical Quality Improvement Program in orthopaedic surgery. Clinical Orthopaedics and Related Research 473:1574–1581, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Mota RE, Tarricone R, Ciani O, Bridges JF, Drummond M. Determinants of demand for total hip and knee arthroplasty: a systematic literature review. BMC Health Services Research 12:225, 2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.National Institute of Health Intraagency Pain Research Coordinating Committee. Federal Pain Research Strategy. Available at: https://iprcc.nih.gov/sites/default/files/iprcc/FPRS_Research_Recommendations_Final_508C.pdf. Accessed December 1, 2018.
- 18.Neuprez A, Neuprez AH, Kaux JF, Kurth W, Daniel C, Thirion T, Huskin JP, Gillet P, Bruyere O, Reginster JY. Total joint replacement improves pain, functional quality of life, and health utilities in patients with late-stage knee and hip osteoarthritis for up to 5 years. Clinical Rheumatology 39:861–871, 2020 [DOI] [PubMed] [Google Scholar]
- 19.Papakostidou I, Dailiana ZH, Papapolychroniou T, Liaropoulos L, Zintzaras E, Karachalios TS, Malizos KN. Factors affecting the quality of life after total knee arthroplasties: a prospective study. BMC Musculoskeletal Disorders. 13:116, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pierce TP, Elmallah RK, Lavernia CJ, Chen AF, Harwin SF, Thomas CM, Mont MA. Racial Disparities in Lower Extremity Arthroplasty Outcomes and Use. Orthopedics 38:e1139–1146, 2015 [DOI] [PubMed] [Google Scholar]
- 21.Pitcher MH, Von Korff M, Bushnell MC, Porter L. Prevalence and Profile of High-Impact Chronic Pain in the United States. J Pain 20:146–160, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Rashiq S, Dick BD. Post-surgical pain syndromes: a review for the non-pain specialist. Canadian Journal of Anaesthesia 61:123–130, 2014 [DOI] [PubMed] [Google Scholar]
- 23.Ravindran D. Chronic postsurgical pain: prevention and management. Journal of Pain & Palliative Care Pharmacotherapy 28:51–53, 2014. [DOI] [PubMed] [Google Scholar]
- 24.Schnabel A, Yahiaoui-Doktor M, Meissner W, Zahn PK, Pogatzki-Zahn EM. Predicting poor postoperative acute pain outcome in adults: an international, multicentre database analysis of risk factors in 50,005 patients. Pain Reports 5:e831, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Schug SA, Lavand’homme P, Barke A, Korwisi B, Rief W, Treede RD. The IASP classification of chronic pain for ICD-11: chronic postsurgical or posttraumatic pain. Pain 160:45–52, 2019 [DOI] [PubMed] [Google Scholar]
- 26.van Dijk GM, Veenhof C, Schellevis F, Hulsmans H, Bakker JP, Arwert H, Dekker JH, Lankhorst GJ, Dekker J. Comorbidity, limitations in activities and pain in patients with osteoarthritis of the hip or knee. BMC Musculoskeletal Disorders 9:95, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.van Dijk GM, Veenhof C, Spreeuwenberg P, Coene N, Burger BJ, van Schaardenburg D, van den Ende CH, Lankhorst GJ, Dekker J, CAPRA Study Group. Prognosis of limitations in activities in osteoarthritis of the hip or knee: a 3-year cohort study. Archives of Physical Medicine and Rehabilitation 91:58–66, 2010 [DOI] [PubMed] [Google Scholar]
- 28.Vandenbroucke JP, von Elm E, Altman DG, Gotzsche PC, Mulrow CD, Pocock SJ, Poole C, Schlesselman JJ, Egger M, STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. Annals of Internal Medicine 147:W163–194, 2007 [DOI] [PubMed] [Google Scholar]
- 29.Von Korff M, DeBar LL, Krebs EE, Kerns RD, Deyo RA, Keefe FJ. Graded chronic pain scale revised: mild, bothersome, and high-impact chronic pain. Pain 161:651–661, 2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Von Korff M, Ormel J, Keefe FJ, Dworkin SF. Grading the severity of chronic pain. Pain 50:133–149, 1992 [DOI] [PubMed] [Google Scholar]
- 31.Wilson R, Pryymachenko Y, Audas R, Abbott JH. Long-term opioid medication use before and after joint replacement surgery in New Zealand. The New Zealand Medical Journal 132:33–47, 2019 [PubMed] [Google Scholar]
- 32.Wylde V, Beswick A, Bruce J, Blom A, Howells N, Gooberman-Hill R. Chronic pain after total knee arthroplasty. EFORT Open Reviews 3:461–470, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Yang MMH, Hartley RL, Leung AA, Ronksley PE, Jette N, Casha S, Riva-Cambrin J. Preoperative predictors of poor acute postoperative pain control: a systematic review and meta-analysis. BMJ Open 9:e025091, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Zhang W, Lyman S, Boutin-Foster C, Parks ML, Pan TJ, Lan A, Ma Y. Racial and Ethnic Disparities in Utilization Rate, Hospital Volume, and Perioperative Outcomes After Total Knee Arthroplasty. The Journal of Bone and Joint Surgery. American Volume 98:1243–1252, 2016 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.