Relationship Between Intraarticular Injections on Patient-reported Outcomes in Total Knee Arthroplasty

Abstract

Introduction:

Total knee arthroplasty (TKA) patients undergo intraarticular (IA) injections before surgery to reduce pain, improve function, and delay surgery. The purpose of this study was to determine the relationship between different types of IA injection exposure and preoperative and postoperative outcomes.

Methods:

We reviewed a series of 321 patients who underwent primary TKA from 2016 to 2022, excluding patients with contralateral TKA within 6 months. Patients were categorized by IA injection type—ketorolac, triamcinolone, or no injection—received within 6 months of TKA. Main outcomes were the Knee Injury and Osteoarthritis Outcome Score (KOOS), and opioid prescriptions received before and 3 months after TKA.

Results:

Most of the 321 patients were female (64%) and White (62%). Of the 321 patients who underwent TKA, 113 (35%) received ketorolac, 64 (20%) received triamcinolone, and 144 (45%) received no injections. Significant differences in preoperative KOOS pain and ADL scores (P = 0.021 and P = 0.047) were observed among groups. Triamcinolone was associated with significantly less preoperative pain (P = 0.016) and greater function (P = 0.046) than ketorolac. No significant differences were found between groups in any KOOS subscale at 3 months post-TKA. Opioid prescriptions, refills, and total morphine milligram equivalents did not differ between injection groups.

Conclusion:

Triamcinolone IA injections were associated with less pain and greater function before TKA than ketorolac. Future studies should include prospective studies, stratified by severity of knee osteoarthritis, to investigate whether exposure to IA injections has a favorable risk-benefit profile, accounting for osteoarthritis disease progression and heterogeneity.

As the US population ages, the number of people with end-stage knee osteoarthritis (OA) who undergo total knee arthroplasty (TKA), the mainstay treatment for this condition, is projected to increase by 673% from 450,000 in 2014 to 3.48 million in 2030.¹ Many people with severe knee OA may delay or refuse TKA for a variety of reasons, such as a desire to avoid the risk of revision surgery in younger patients, fear of postoperative pain and complications, and interference with work.^2,3 Nonsurgical treatment options available to patients with severe knee OA include pharmacological therapies, cryotherapy, physical therapy, and intraarticular (IA) injections. Up to 90% of patients with knee OA have received at least one IA injection before TKA, with some patients receiving up to four IA injections annually and others receiving a single injection.^4,5

Given the widespread use of IA injections in patients with knee OA, it is important to understand the risks and benefits of this treatment in patients who subsequently undergo TKA. Recent large-scale studies have found that IA corticosteroid and hyaluronic acid (HA) injections given within 3 months before TKA are associated with an increased risk of periprosthetic joint infection (PJI).^6-12 Currently, most clinicians avoid administering IA steroids within 90 days of TKA to reduce the risk of PJI.¹³ In a study examining the relationship between preoperative IA injection exposure and chronic opioid use, a single corticosteroid IA injection received within 1 year of TKA decreased the risk of chronic opioid use after surgery, whereas two or more corticosteroid injections increased this risk.¹⁴ By contrast, TKA patients with two or more HA injections in the year before surgery had lower odds of chronic opioid use.¹²

The effects of exposure to any preoperative IA knee injection, and exposure to different IA-administered drugs, on patient-reported outcomes (PROs) before and after TKA have not been fully explored. One study found that the number of IA injections patients received before TKA was not associated with increased rates of complications, PJI, or poor short-term functional outcomes (a decline in the Knee Society Score) 12 months after TKA.⁵ A randomized, placebo-controlled trial found that receiving a single IA methylprednisolone acetate injection 1 week before TKA did not reduce acute postoperative pain or opioid use during the first 14 days after surgery despite causing a notable decrease in IA interleukin-6 in patients with preoperative high-pain knee OA and central sensitization.^6-11,15 This study is intended to address the unmet need for more information regarding the clinical effect of IA injections on TKA outcomes. The primary purpose of this study was to investigate whether corticosteroid and ketorolac IA injections received during the 6 months before TKA are associated with preoperative and postoperative PROs. In addition, we will examine if injection type is related to opioid prescriptions after TKA.

Methods

The Institutional Review Board approved this retrospective chart review and waived informed consent because the study used deidentified patient data collected during routine care. Eligible patients were adults aged ≥18 years who underwent unilateral primary TKA done by a single surgeon between 2016 and 2022. Patients who had a contralateral TKA within 6 months after primary TKA were excluded.

Data collected from patients' charts included patient demographics (sex, age, race, insurance type, and surgery date), body mass index, Kellgren-Lawrence grade, Knee Injury and Osteoarthritis Outcome Score (KOOS), and history of contralateral TKA.

Opioid prescription data were collected during the 3 months before and after TKA using the Louisiana Prescription Monitoring Program database. These data were used to derive morphine milligram equivalents (MME) and the number of refilled prescriptions.

Pain Protocol

All patients received the same preoperative multimodal pain protocol consisting of percutaneous cryoneurolysis, 150 mg pregabalin, 200 mg celecoxib, and 1,000 mg intravenous acetaminophen. Immediately before surgery, an anesthesiologist administered spinal anesthesia with 1.6 mg of 0.75% bupivacaine and occasional intravenous fentanyl based on provider preference, in addition to an adductor canal block using liposomal bupivacaine (Exparel; Pacira Pharmaceuticals). A periarticular infiltration of 0.25% bupivacaine hydrochloride (Marcaine; Pfizer) was given intraoperatively. All patients treated before May 1, 2020, also received an opioid prescription before discharge. Beginning on May 1, 2020, opioid-naïve patients were only prescribed opioids at or after discharge on patient request after discharge. Groups were divided into “automatic” and “upon request” based on differences in opioid prescribing practices. Opioid use data were not collected on patients who underwent TKA in 2016 because the Prescription Monitoring Program database was not yet established.

Measures

KOOS was assessed at the preoperative visit before TKA and 3 months after TKA. The KOOS has five subscales, four of which were of interest in this study: pain, symptoms, function in daily living (ADL), and quality of life.¹⁵ Subscale scores were transformed to a 0 to 100 scale, with 0 representing extreme problems and 100 representing no problems.¹⁵ Studies have reported variable minimal clinically important differences (MCIDs) for the KOOS based on different patient cohorts, interventions, follow-up periods, and calculation methods.¹⁶ A systematic review of TKA studies reported a median MCID of 12 for KOOS pain and KOOS ADL.¹⁶ Several more recently conducted studies have reported lower thresholds for the KOOS pain subscale MCID (5 and 7.9) in TKA patients.^17,18 ICD-10 (10th revision of the International Classification of Diseases) codes were used to identify IA injection procedures done in the 6 months before TKA and chart review to identify the injected drug (triamcinolone or ketorolac). Too few patients had received HA IA injections to be included in analyses. For patients who received multiple IA injections within the study period, the injection drug was characterized by the last IA injection received before TKA.

Opioid prescriptions received before and after TKA were compared with IA injection type (no injection, triamcinolone injection, or ketorolac injection). In the 3-month period before and after TKA, opioid use was defined as filling at least one opioid prescription; the number of refills and total MME were also calculated.

Statistical Analysis

Study data were collected and managed using REDCap electronic data capture.^19,20 Data were analyzed with SAS version 9.4 (SAS Institute). Baseline characteristics were compared with the IA injection group (none, triamcinolone, or ketorolac) using the chi square test for categorical variables with cell counts greater than five, Fisher exact test for categorical variables with cell counts less than five, and Student t-test for continuous normally distributed variables. Opioid prescriptions during the 3 months before and after TKA were compared with IA injection group using the exact test for binary outcomes and the Kruskal-Wallis test for medians. PROs before TKA were compared with the IA injection group using analysis of variance and PROs assessed 3 months post-TKA were compared using analysis of covariance adjusting for preoperative PROs and sex. When a notable injection group effect was observed, the least square means were compared using the Tukey multiple comparison adjustment. Furthermore, a subanalysis was conducted including only patients who received any injection(s) before TKA, with the number of injections included as a covariate. Assumptions of normally distributed, independent, and identically distributed residuals with homogeneous variances were met in applicable parametric models. Statistical significance was set at a two-sided alpha of less than 0.05.

Results

Patient Characteristics

A total of 321 patients were included in this study, of whom 144 (44.9%) received no IA injections, 113 (35.2%) received at least one ketorolac IA injection, and 64 (19.9%) received at least one triamcinolone IA injection during the 6 months before TKA. Last ketorolac and triamcinolone IA injections were administered at a median (interquartile range) of 69 (48, 116) days and 83 (53, 125) days before surgery, respectively, and did not significantly differ (P = 0.335).

Table 1 describes the demographics, radiographic severity of knee OA, and body mass index by IA treatment. There were no statistically significant differences between groups. The study sample was predominately female (68.5%) with severe radiographic knee OA (KL grade 4, 87.9%) and an average age of 68 years. The sample included 192 (59.8%) White patients, 118 (36.8%) Black patients, and 11 (3.4%) patients who were categorized as “other race.” Most (64.5%, 207/321) patients had private insurance or Medicaid advantage, 23.1% (74/321) had Medicare, and 10.3% (33/321) were Medicaid beneficiaries.

Table 1.

Patient Demographics and Clinical Characteristics by Injection Type (N = 321)

Characteristic	None (n = 144)	KET (n = 113)	TRI (n = 64)	P Value
Sex, % (n)				0.096
Male	36.1 (52)	23.9 (27)	34.4 (22)
Female	63.9 (92)	76.1 (86)	65.6 (42)
Race, % (n)				0.848
Black or African American	34.0 (49)	38.9 (44)	39.1 (25)
White or Caucasian	61.8 (89)	57.5 (65)	59.4 (38)
Other	4.2 (6)	3.6 (4)	1.5 (1)
Insurance type, % (n)				0.739
Private	37.5 (54)	34.5 (39)	40.7 (26)
Medicare	27.1 (39)	19.5 (22)	20.3 (13)
Medicaid	9.7 (14)	12.4 (14)	7.8 (5)
Medicare advantage	24.3 (35)	31.0 (35)	28.1 (18)
Others	1.4 (2)	2.6 (3)	3.1 (2)
Kellgren-Lawrence grade, % (n)				0.423
0	0 (0)	0 (0)	0 (0)
1	0 (0)	0 (0)	0 (0)
2	0.7 (1)	0 (0)	3.1 (2)
3	10.4 (15)	12.4 (14)	10.9 (7)
4	88.9 (128)	87.6 (99)	85.9 (55)
Age in yr, mean (SD)	67.9 (9.2)	67.8 (10.0)	68.2 (8.0)	0.967
Body mass index in kg/m2, mean (SD)	32.0 (5.7)	32.5 (5.9)	33.3 (6.5)	0.338

KET = ketorolac, TRI = triamcinolone

Knee Injury and Osteoarthritis Outcome Score

Table 2 reports KOOS scores before TKA and 3 months after TKA by IA treatment group (higher KOOS scores indicate better outcomes). Significant differences were observed among the three groups in preoperative KOOS pain (P = 0.021) and KOOS ADL (P = 0.047) subscale scores, but not for KOOS symptoms (P = 0.093) and KOOS quality of life (P = 0.236). Triamcinolone was associated with significantly worse preoperative KOOS pain (P = 0.016) and ADL (P = 0.046) scores compared with the ketorolac group. No notable differences were observed in post-TKA KOOS subscale scores between groups. A post hoc analysis of patients who received any injection(s) within 6 months of TKA adjusting for the number of injections received yielded similar results (Supplemental Table A, http://links.lww.com/JG9/A490).

Table 2.

Knee Injury and Osteoarthritis Outcome Score Subscales Before and 3 Months After Total Knee Arthroplasty by Type of Intraarticular Injection (N = 321)

Item	None (n = 144)	KET (n = 113)	TRI (n = 64)	P Value
Before TKA, mean (SD)
Pain	36.9* (18.9)	35.0# (18.9)	43.4** (18.8)	0.021
Symptoms	40.0 (20.8)	38.4 (19.9)	45.3 (19.2)	0.093
ADL	38.7* (20.0)	37.8# (19.8)	45.4** (18.8)	0.047
QOL	20.3 (17.5)	20.9 (18.2)	25.0 (20.0)	0.236
3 mo after TKA, LSM (SEM, n)1
Pain	68.1 (2.3, 77)	66.6 (2.5, 64)	65.0 (3.3, 36)	0.729
Symptoms	65.4 (2.1, 78)	64.3 (2.3, 64)	63.1 (3.0, 38)	0.802
ADL	73.6 (2.1, 77)	70.0 (2.3, 64)	67.6 (3.0, 37)	0.199
QOL	51.2 (2.7, 78)	50.2 (3.1, 62)	48.1 (4.0, 36)	0.819

ADL = function in daily living, KET = ketorolac, LSM = least square means, QOL = quality of life, SEM = standard error of the mean, TKA = total knee arthroplasty, TRI = triamcinolone

^#Means with different superscripts are significantly different from each other at P < 0.05.

¹Analysis included before TKA value and sex as covariates.

Opioid Prescriptions

Table 3 displays the association between IA treatment and opioids prescribed 3 months before and after TKA. No notable differences were observed by IA treatment in the prevalence of preoperative and postoperative opioid prescriptions or refills received, and total MME before and after surgery. Analyses also revealed no notable differences for all post-TKA opioid variables within the automatic and by-request subgroups.

Table 3.

Opioid Prescriptions 3 Months Before and After Total Knee Arthroplasty by Intraarticular Injection Treatment (N = 274)^a

Outcome	None (n = 121)	KET (n = 105)	TRI (n = 48)	P Value
Before TKA
≥1 filled opioid prescription, % (n)	22.3 (27)	17.1 (18)	33.3 (16)	0.090
MME over 3 mo, median (IQR)	600 (75, 1800)	585 (200, 2025)	300 (150, 1405)	0.836
After TKAb
Automatic opioids at discharge	n = 58	n = 56	n = 42
≥1 filled opioid prescription, % (n)	96.6 (56)	98.2 (55)	97.6 (41)	0.851
≥1 refills, % (n)	50.0 (29)	66.1 (37)	54.6 (24)	0.214
MME over 3 mo, median (IQR)	593 (308, 1530)	765 (315, 1410)	660 (315, 1125)	0.617
On request opioids at discharge	n = 63	n = 49	n = 6
≥1 filled opioid prescription, % (n)	38.1 (24)	53.1 (26)	33.3 (2)	0.241
≥1 refills, % (n)	19.1 (12)	22.5 (11)	16.7 (1)	0.926
MME over 3 mo, median (IQR)	368 (218, 1050)	280 (140, 945)	740 (280, 1200)	0.822

IQR = interquartile range, KET = ketorolac, MME = morphine milligram equivalent, TKA = total knee arthroplasty, TRI = triamcinolone

^aOpioid data were not collected on patients who had a TKA in 2016 (n = 47).

^bBefore May 1, 2020, all patients received an automatic opioid prescription at discharge; after May 1, 2020, all discharge opioid prescriptions were provided only on request.

Discussion

The results from this study indicate that patients treated with triamcinolone IA injections had markedly less pain and improved function before TKA than patients treated with ketorolac injections. Several studies have compared the effectiveness of ketorolac and corticosteroid injections in reducing pain and improving function in patients with knee OA, producing conflicting results.^21-24 Currently, there is no consensus regarding which injection type may be optimal for improving pain and function. We found that patients who received triamcinolone reported statistically significantly less knee pain preoperatively than those who received ketorolac. The difference in preoperative KOOS pain between triamcinolone and ketorolac was 8.3 points, which exceeds the MCIDs of five and 7.9, reported in recent studies,^18,25 but falls short of the median MCID of 12 reported in a recent systematic literature review¹⁶; therefore, it is questionable whether the differences between preoperative KOOS pain and ADL scores between IA groups are clinically meaningful.

Other extrinsic factors not captured in PROs can affect the MCID such as complications and cost which require consideration. Although our data suggest that triamcinolone is more effective in improving pain before TKA, cost should be considered: IA ketorolac is 143% less expensive per injection than corticosteroids.²³ Over 50% (162/321) of patients included in our study were insured by Medicare or Medicare Advantage, which may not always fully cover injections.

In addition to financial implications, physicians must weigh the risks associated with IA injections for their patients. A recent study indicated that 7% to 8% of patients receiving corticosteroid-based IA injections may experience accelerated disease progression.²⁶ Although rare, patients may also experience subchondral bone collapse, osteonecrosis, joint infection, skin atrophy, and tendinopathy.^22,26 Ketorolac, a nonsteroidal anti-inflammatory drug, is associated with several drug-drug interactions and has been shown to increase the risk of gastrointestinal toxicity, chronic renal disease, and cardiovascular thrombotic events.^27-29 Although injections may provide short-term pain relief, our results show that this effect is negligible 3 months after surgery. Physicians in hospital settings will increasingly need to decide whether to administer injections that may provide temporary relief but carry notable side effects. Although triamcinolone injections may provide more preoperative pain relief, there was little effect on pain differences between injection and noninjection groups after TKA. This is consistent with previous work showing that IA corticosteroid injections do not reduce pain in the acute period after TKA in patients with high pain levels.³⁰

There were no differences in opioid use between injection groups before or after TKA. Patients are likely experiencing similar amounts of postoperative pain regardless of injection type received, which could account for the lack of difference in opioid use across groups. This is further supported by the lack of differences observed in KOOS pain scores after TKA.

This study has several limitations. First, the study's retrospective design and reliance on database inputs precludes conclusions about causality and limited the ability to control for potentially confounding factors. Second, because all cases done by a single orthopaedic surgeon at a single site, generalizability of results is limited. Third, the duration of pain relief may vary between injection types and among patients, as studies have shown that effects could last between 1 and 3 months.^31,32 Finally, only a very small number of patients received IA HA, which made us unable to study this injection type because of insufficient power.

Conclusion

Although our results indicate statistically significant differences in preoperative KOOS Pain and ADL subscale scores, it is likely that these differences constitute minimal clinical differences between groups before surgery, and these differences appear to dissipate entirely after TKA. The lack of a well-defined improvement in postoperative PROs casts uncertainty on the practicality of preoperative injection use in patients with end-stage OA with TKA likely imminent. However, the decision to discontinue preoperative injections is made more difficult when faced with patient demand for pain relief in the clinical setting. Therefore, it is essential for physicians and patients with end-stage knee OA to consider the risks and benefits of IA injections carefully.

Supplementary Material

jagrr-10-e25.00064-s001.docx ^{(15.4KB, docx)}