Intra-articular corticosteroids and the risk of knee osteoarthritis progression: results from the Osteoarthritis Initiative

  • C. Zeng
    Affiliations
    Department of Orthopaedics, Xiangya Hospital, Central South University, Changsha, Hunan, China

    Division of Rheumatology, Allergy, and Immunology, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
    Search for articles by this author
  • N.E. Lane
    Affiliations
    Center for Musculoskeletal Health and Department of Medicine, University of California School of Medicine, Sacramento, CA, USA
    Search for articles by this author
  • D.J. Hunter
    Affiliations
    Rheumatology Department, Royal North Shore Hospital and Institute of Bone and Joint Research, Kolling Institute, University of Sydney, Sydney, Australia
    Search for articles by this author
  • J. Wei
    Affiliations
    Division of Rheumatology, Allergy, and Immunology, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA

    Health Management Center, Xiangya Hospital, Central South University, Changsha, Hunan, China
    Search for articles by this author
  • H.K. Choi
    Affiliations
    Division of Rheumatology, Allergy, and Immunology, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
    Search for articles by this author
  • T.E. McAlindon
    Affiliations
    Division of Rheumatology, Tufts Medical Center, Boston, MA, USA
    Search for articles by this author
  • H. Li
    Affiliations
    Department of Orthopaedics, Xiangya Hospital, Central South University, Changsha, Hunan, China
    Search for articles by this author
  • N. Lu
    Affiliations
    Division of Rheumatology, Allergy, and Immunology, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
    Search for articles by this author
  • G. Lei
    Correspondence
    Address correspondence and reprint requests to: G. Lei, Department of Orthopaedics, Xiangya Hospital, Central South University, Changsha, Hunan, China.
    Affiliations
    Department of Orthopaedics, Xiangya Hospital, Central South University, Changsha, Hunan, China

    Hunan Key Laboratory of Joint Degeneration and Injury, Changsha, Hunan, China

    National Clinical Research Center of Geriatric Disorders, Xiangya Hospital, Central South University, Changsha, Hunan, China
    Search for articles by this author
  • Y. Zhang
    Correspondence
    Address correspondence and reprint requests to: Y. Zhang, Division of Rheumatology, Allergy, and Immunology, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA.
    Affiliations
    Division of Rheumatology, Allergy, and Immunology, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
    Search for articles by this author
Open ArchivePublished:January 28, 2019DOI:https://doi.org/10.1016/j.joca.2019.01.007

      Summary

      Objective

      A recent randomized clinical trial reported that repeated intra-articular corticosteroids (IACs) were associated with a greater cartilage loss. This study aimed to examine the relation of IACs to knee radiographic osteoarthritis (ROA) progression in a real-world setting.

      Design

      A cohort that initiated IACs and a comparison cohort without IACs from participants with mild to moderate knee ROA in the Osteoarthritis Initiative (OAI) were assembled (from 0-month to 48-month). Two measures of knee ROA progression were assessed during the follow-up period: (1) an increase in Kellgren and Lawrence (KL) grade by ≥1 grade or having a knee replacement (i.e., KL grade worsening); and (2) a decrease in joint space width (JSW) by ≥0.7 mm or having a knee replacement (i.e., JSW worsening). The associations of IACs initiation using a propensity-score matched cohort study and continuous IACs using marginal structural models with the risk of knee ROA progression were examined.

      Results

      Among 684 propensity-score matched participants at baseline (148 IACs initiators, 536 comparators), 65 knees (21.7/100 person-years) in the IACs initiation cohort and 90 knees (7.1/100 person-years) in the comparison cohort experienced KL worsening. The hazard ratios (HRs) of KL worsening from IACs initiation and continuous IACs were 3.02 (95% confidence interval [CI], 2.19–4.16) and 4.67 (95% CI, 2.92–7.47), respectively. The corresponding HRs of JSW worsening were 2.93 (95% CI, 2.13–4.02) and 3.26 (95% CI, 1.78–5.96), respectively. All HRs for continuous use of IACs were further away from the null.

      Conclusions

      IACs, especially continuous IACs, may be associated with an increased risk of knee ROA progression.

      Keywords

      Introduction

      Osteoarthritis (OA) is the most common joint disorder among older adults
      • GBD 2016 DALYs and HALE Collaborators
      Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016.
      . Pain from OA is a key symptom in the decision to seek medical care and an antecedent to disability. To date few safe and effective treatments for OA are available, and the main goal of clinical management remains the control of pain and improvement of both function and quality of life with the avoidance of therapeutic toxicity
      • Glyn-Jones S.
      • Palmer A.J.
      • Agricola R.
      • Price A.J.
      • Vincent T.L.
      • Weinans H.
      • et al.
      Osteoarthritis.
      .
      Intra-articular corticosteroids (IACs) are a frequently-used treatment regimen for pain relief from symptomatic knee OA as it inhibits inflammation
      • McCabe P.S.
      • Parkes M.J.
      • Maricar N.
      • Hutchinson C.E.
      • Freemont A.
      • O'Neill T.W.
      • et al.
      Brief report: synovial fluid white blood cell count in knee osteoarthritis: association with structural findings and treatment response.
      • Hall M.
      • Doherty S.
      • Courtney P.
      • Latief K.
      • Zhang W.
      • Doherty M.
      Ultrasound detected synovial change and pain response following intra-articular injection of corticosteroid and a placebo in symptomatic osteoarthritic knees: a pilot study.
      and reduces prostaglandin synthesis
      • O'Banion M.K.
      • Winn V.D.
      • Young D.A.
      cDNA cloning and functional activity of a glucocorticoid-regulated inflammatory cyclooxygenase.
      • Rhen T.
      • Cidlowski J.A.
      Antiinflammatory action of glucocorticoids– new mechanisms for old drugs.
      . While meta-analyses have reported IACs may be beneficial in relieving pain, at least for a short period of time (up to 6 weeks after IACs)
      • Juni P.
      • Hari R.
      • Rutjes A.W.
      • Fischer R.
      • Silletta M.G.
      • Reichenbach S.
      • et al.
      Intra-articular corticosteroid for knee osteoarthritis.
      • Bannuru R.R.
      • Schmid C.H.
      • Kent D.M.
      • Vaysbrot E.E.
      • Wong J.B.
      • McAlindon T.E.
      Comparative effectiveness of pharmacologic interventions for knee osteoarthritis: a systematic review and network meta-analysis.
      , recommendations of its use in management of knee OA varies. For example, the Osteoarthritis Research Society International (OARSI), the American College of Rheumatology (ACR), and the National Institute for Health and Care Excellence (NICE) recommended or conditionally recommended IACs for patients with knee OA
      • Hochberg M.C.
      • Altman R.D.
      • April K.T.
      • Benkhalti M.
      • Guyatt G.
      • McGowan J.
      • et al.
      American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee.
      • McAlindon T.E.
      • Bannuru R.R.
      • Sullivan M.C.
      • Arden N.K.
      • Berenbaum F.
      • Bierma-Zeinstra S.M.
      • et al.
      OARSI guidelines for the non-surgical management of knee osteoarthritis.
      • National Clinical Guideline Centre
      Osteoarthritis: Care and Management in Adults.
      ; however, the American Academy of Orthopaedic Surgeons (AAOS) did not reach a conclusion on its use for the management of knee OA
      • Jevsevar D.S.
      Treatment of osteoarthritis of the knee: evidence-based guideline, 2nd edition.
      .
      While previous in vivo studies have reported a detrimental effect of IACs on cartilage
      • Behrens F.
      • Shepard N.
      • Mitchell N.
      Alterations of rabbit articular cartilage by intra-articular injections of glucocorticoids.
      • Mankin H.J.
      • Conger K.A.
      The acute effects of intra-articular hydrocortisone on articular cartilage in rabbits.
      • Murray R.C.
      • DeBowes R.M.
      • Gaughan E.M.
      • Zhu C.F.
      • Athanasiou K.A.
      The effects of intra-articular methylprednisolone and exercise on the mechanical properties of articular cartilage in the horse.
      , there is a paucity of data of the effect of IACs on joint structure changes among patients with knee OA. Results from two randomized controlled trials (RCTs) assessing effect of repeated IACs on risk of knee structure change are conflicting
      • Raynauld J.P.
      • Buckland-Wright C.
      • Ward R.
      • Choquette D.
      • Haraoui B.
      • Martel-Pelletier J.
      • et al.
      Safety and efficacy of long-term intraarticular steroid injections in osteoarthritis of the knee: a randomized, double-blind, placebo-controlled trial.
      • McAlindon T.E.
      • LaValley M.P.
      • Harvey W.F.
      • Price L.L.
      • Driban J.B.
      • Zhang M.
      • et al.
      Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial.
      . One trial conducted in the early 2000's reported that repeated IACs were not associated with change in joint space width (JSW) compared with intra-articular saline
      • Raynauld J.P.
      • Buckland-Wright C.
      • Ward R.
      • Choquette D.
      • Haraoui B.
      • Martel-Pelletier J.
      • et al.
      Safety and efficacy of long-term intraarticular steroid injections in osteoarthritis of the knee: a randomized, double-blind, placebo-controlled trial.
      ; and another trial conducted recently showed that repeated IACs resulted in greater loss of cartilage volume, albeit the amount of cartilage loss was small and may not be clinically meaningful, compared with intra-articular saline
      • McAlindon T.E.
      • LaValley M.P.
      • Harvey W.F.
      • Price L.L.
      • Driban J.B.
      • Zhang M.
      • et al.
      Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial.
      . These findings have led clinicians to question IACs' utility in the management of knee OA
      • Bliddal H.
      • Henriksen M.
      Osteoarthritis: time to put steroid injections behind us?.
      . To address this knowledge gap, we examined the relation of IACs to knee radiographic osteoarthritis (ROA) progression using data collected from the Osteoarthritis Initiative (OAI).

      Methods

       Study population

      The OAI is a multi-center longitudinal observational study of risk factors for both incident and progressive knee OA. A detailed description regarding the rationale and approach of the OAI can be found at https://data-archive.nimh.nih.gov/oai/about-oai. In the current analyses, we used data collected from the 0-, 12-, 24-, 36-, and 48-month visits where the assessments of knee radiographs for progressive cohorts are publicly available.

       Assessment of IACs

      At 0-month and each annual follow-up visit, participants were asked “During the past 6 months, have you had an injection of steroids (cortisone, corticosteroids) in either of your knees for treatment of your arthritis?” If they answered “yes”, participants were then queried about in which knee the injection was given. The visit when the participant reported initiation of IACs was deemed as the “index visit”.

       Assessment of ROA progression endpoints

      Knee ROA progression was assessed using two endpoints: (1) Kellgren and Lawrence (KL) grade worsening; and (2) JSW worsening. Specifically, bilateral knee radiographs were taken at 0-month and each annual follow-up visit. KL grade at the tibiofemoral joint was assessed at the central reading center. Among knees with KL grade 2 or 3 at baseline (i.e., the nearest visit prior to the “index visit”), we defined a knee as KL grade worsening if its KL grade at a follow-up visit had increased by ≥1 grade compared with that at baseline or the patient had received a knee replacement. As the vast majority of patients’ tibiofemoral OA occur at the medial compartment, our primary outcome regarding joint space narrowing was medial tibiofemoral JSW assessed at a fixed location along the tibial plateaus (i.e., 0.250) using a semi-automated method
      • Duryea J.
      • Neumann G.
      • Niu J.
      • Totterman S.
      • Tamez J.
      • Dabrowski C.
      • et al.
      Comparison of radiographic joint space width with magnetic resonance imaging cartilage morphometry: analysis of longitudinal data from the osteoarthritis initiative.
      . Two expert readers with over 50 years of combined experience independently assessed the medial tibiofemoral JSW (intraclass correlation coefficient [ICC] = 0.98 for cross-sectional analysis; ICC = 0.94 for 3-year longitudinal analysis), blinded to clinical data
      • Ratzlaff C.
      • Ashbeck E.L.
      • Guermazi A.
      • Roemer F.W.
      • Duryea J.
      • Kwoh C.K.
      A quantitative metric for knee osteoarthritis: reference values of joint space loss.
      . We defined a knee as having JSW worsening if its value at a follow-up visit decreased by ≥0.7 mm from baseline
      • Ornetti P.
      • Brandt K.
      • Hellio-Le Graverand M.P.
      • Hochberg M.
      • Hunter D.J.
      • Kloppenburg M.
      • et al.
      OARSI-OMERACT definition of relevant radiological progression in hip/knee osteoarthritis.
      or the patient had knee replacement during the follow-up.

       Selection of index knee

      We used an index knee from each participant, and selection of the index knee for KL grade or JSW worsening in the current analysis is depicted in Fig. 1. Briefly, eligible participants consisted of those who had at least one knee with a KL grade of 2 or 3 and did not report IACs use at baseline
      • Choi H.K.
      • Nguyen U.S.
      • Niu J.
      • Danaei G.
      • Zhang Y.
      Selection bias in rheumatic disease research.
      . For JSW worsening, knees were ineligible if they had definite lateral joint space narrowing >1 on an ordinal 0 (i.e., normal joint width) to 3 scale according to the OARSI grades
      • Altman R.D.
      • Gold G.E.
      Atlas of individual radiographic features in osteoarthritis, revised.
      • Eckstein F.
      • Wirth W.
      • Nevitt M.C.
      Recent advances in osteoarthritis imaging–the osteoarthritis initiative.
      as including those with lateral joint space narrowing may incorporate knees that have pseudo-widening (a consequence of increasing lateral joint space narrowing), or minimal tibial plateau-rim distance ≥6.5 mm or missing
      • Eckstein F.
      • Collins J.E.
      • Nevitt M.C.
      • Lynch J.A.
      • Kraus V.B.
      • Katz J.N.
      • et al.
      Brief report: cartilage thickness change as an imaging biomarker of knee osteoarthritis progression: data from the foundation for the national Institutes of Health osteoarthritis biomarkers consortium.
      • Ruhdorfer A.
      • Haniel F.
      • Petersohn T.
      • Dorrenberg J.
      • Wirth W.
      • Dannhauer T.
      • et al.
      Between-group differences in infra-patellar fat pad size and signal in symptomatic and radiographic progression of knee osteoarthritis vs non-progressive controls and healthy knees - data from the FNIH biomarkers consortium study and the osteoarthritis initiative.
      , or tibial-plateau-rim distance >2 mm of the minimal distance of all visits
      • Eckstein F.
      • Collins J.E.
      • Nevitt M.C.
      • Lynch J.A.
      • Kraus V.B.
      • Katz J.N.
      • et al.
      Brief report: cartilage thickness change as an imaging biomarker of knee osteoarthritis progression: data from the foundation for the national Institutes of Health osteoarthritis biomarkers consortium.
      • Ruhdorfer A.
      • Haniel F.
      • Petersohn T.
      • Dorrenberg J.
      • Wirth W.
      • Dannhauer T.
      • et al.
      Between-group differences in infra-patellar fat pad size and signal in symptomatic and radiographic progression of knee osteoarthritis vs non-progressive controls and healthy knees - data from the FNIH biomarkers consortium study and the osteoarthritis initiative.
      or baseline JSW <0.7 mm. If a subject had two eligible knees, the one with KL grade 3 was selected as the index knee; if the KL grade for both knees was equal, then the one with the higher Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) knee pain score was selected as the index knee; if the WOMAC pain score for both knees was equal, we randomly selected one as the index knee.
      Fig. 1
      Fig. 1Flow Chart of Participants' Selection Process for the Outcome of KL (Left) or JSW (Right) Worsening. OAI, Osteoarthritis Initiative; KL, Kellgren and Lawrence; JSW, joint space width; IACs, intra-articular corticosteroids. Each subject only contributed one knee for the current analysis. If a subject had two eligible knees, the one with KL grade 3 was selected as the index knee; if the KL grade for both knees was equal, then the one with the higher Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) knee pain score was selected as the index knee; if the WOMAC pain score for both knees was equal, we randomly selected one as the index knee.

       Assessment of covariates

      Age, sex, race, education, history of knee injury (defined as knee injured badly enough to limit ability to walk for at least 2 days), Physical Activity Scale for Elderly (PASE), non-steroidal anti-inflammatory drugs (NSAIDs), WOMAC knee pain, Charlson Comorbidity Index (CCI), and widespread pain
      • Carlesso L.C.
      • Niu J.
      • Segal N.A.
      • Frey-Law L.A.
      • Lewis C.E.
      • Nevitt M.C.
      • et al.
      The effect of widespread pain on knee pain worsening, incident knee osteoarthritis (OA), and incident knee pain: the multicenter OA (MOST) study.
      were collected at baseline. Body mass index (BMI) was computed as weight (kg)/height (m)
      • Glyn-Jones S.
      • Palmer A.J.
      • Agricola R.
      • Price A.J.
      • Vincent T.L.
      • Weinans H.
      • et al.
      Osteoarthritis.
      . In each annual follow-up visit, data were also collected on body weight, WOMAC knee pain, knee injury, NSAIDs use, PASE, and widespread pain.

       Statistical analysis

      For each 1-year time block, we calculated a propensity score for each knee with the initiation of IACs (i.e., a conditional probability of initiating IACs by accounting for a vector of observed covariates
      • Rosenbaum P.R.
      • Rubin D.B.
      The central role of the propensity score in observational studies for causal effects.
      ) using logistic regression
      • Lu N.
      • Misra D.
      • Neogi T.
      • Choi H.K.
      • Zhang Y.
      Total joint arthroplasty and the risk of myocardial infarction: a general population, propensity score-matched cohort study.
      • Seeger J.D.
      • Williams P.L.
      • Walker A.M.
      An application of propensity score matching using claims data.
      . Variables included in the model consisted of age, sex, race, education, BMI, history of knee injury, PASE, NSAIDs use, WOMAC knee pain score, CCI, widespread pain, and KL grade (for KL worsening) or JSW value (for JSW worsening) assessed at the visit prior to initiation of IACs. Within each time block, for each knee that initiated IACs (i.e., knee in the IACs initiation cohort) we identified up to four propensity-score matched knees that did not receive IACs (i.e., knees in the comparison cohort), and assigned the no-IACs knees with the same “index visit” as their matched knees. The absolute allowable difference in propensity score among knees in each matched set was set to be <0.005. The balance of the covariates between the IAC initiation and comparison cohorts was assessed by standardized differences
      • Austin P.C.
      Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies.
      • Franklin J.M.
      • Rassen J.A.
      • Ackermann D.
      • Bartels D.B.
      • Schneeweiss S.
      Metrics for covariate balance in cohort studies of causal effects.
      .
      First, we took intention-to-treat approach to examine the relation of IACs initiation to the risk of KL (JSW) worsening, assuming participants who initiated IACs remained in that group until they had outcomes or were censored. We calculated the incidence rate for KL (or JSW) worsening and plotted a cumulative incidence curve of each of these outcomes for the IACs cohort and its comparison cohort, respectively. We used cause-specific Cox proportional hazards models to estimate the hazard ratio (HR) and its 95% confidence interval (CI) for the risk of KL (or JSW) worsening for IACs adjusting for competing event (i.e., death). We tested the proportional hazards assumption for each comparison cohort using the Kolmogorov supremum test
      • Lin D.Y.
      • Wei L.-J.
      • Ying Z.
      Checking the Cox model with cumulative sums of martingale-based residuals.
      . If the proportional hazard assumption was violated, we conducted a weighted Cox regression to obtain a non-proportional HR
      • Dunkler D.
      • Ploner M.
      • Schemper M.
      • Heinze G.
      Weighted Cox regression using the R package coxphw.
      . Second, since participants may report IACs during some of the annual visits but not in others, to account for time-varying confounders, we performed a marginal structural model to assess the effect of continuous IACs (versus no-IACs) and the risk of KL (or JSW) worsening
      • Hernan M.A.
      • Brumback B.
      • Robins J.M.
      Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men.
      • Toh S.
      • Hernandez-Diaz S.
      • Logan R.
      • Rossouw J.E.
      • Hernan M.A.
      Coronary heart disease in postmenopausal recipients of estrogen plus progestin therapy: does the increased risk ever disappear? A randomized trial.
      . Specifically, we fitted pooled logistic models to obtain their predicted values for each person-year remaining off IACs and uncensored. We then used a SAS data step to calculate the stabilized inverse-probability weights for each person-year from the predicted values of the previous models. Last, we used generalized estimating equations to fit the final weighted pooled logistic model that estimates the causal parameter of IACs and its robust standard error. Variables in the calculation of the propensity score were included in these models. The marginal structural models estimate the average adherence-adjusted HR of ROA progression for continuous IACs vs no-IACs during the study period
      • Toh S.
      • Hernandez-Diaz S.
      • Logan R.
      • Rossouw J.E.
      • Hernan M.A.
      Coronary heart disease in postmenopausal recipients of estrogen plus progestin therapy: does the increased risk ever disappear? A randomized trial.
      .
      Since patients with more severe knee OA are more likely to initiate and continuously receive IACs as well as are at a higher risk of ROA progression than those with less severe disease, we conducted sensitivity analyses to assess the potential impact of residual confounding. First, we applied asymmetric propensity-score trimming to exclude patients who were receiving IACs most contrary to their prediction, i.e., excluding knees with propensity scores <2.5% based on its distribution in IACs cohort and >97.5% based on its distribution in comparison cohort, respectively
      • Sturmer T.
      • Rothman K.J.
      • Avorn J.
      • Glynn R.J.
      Treatment effects in the presence of unmeasured confounding: dealing with observations in the tails of the propensity score distribution–a simulation study.
      . Second, since either propensity-score matched study design or marginal structural models may not account for unmeasured confounders, we performed quantitative sensitivity analyses to assess the minimum association that an unmeasured confounder(s) would need to have with both IACs and knee ROA progression conditional on the measured covariates in order to explain away an association observed in the current analyses
      • VanderWeele T.J.
      • Ding P.
      Sensitivity analysis in observational research: introducing the E-value.
      . Third, considering that some knees with replacement therapy may not experience KL grade (or JSW) worsening according to the criteria in the current analyses, we repeated these analyses by treating knee replacement as a competing event. Fourth, to minimize the confounding by indication we calculated the propensity score by additionally adding KL grade change (or JSW change) within 12 months before the index visit into logistic regression model. Finally, to minimize potential bias of reverse causality we performed a time-lag analysis by excluding participants with KL grade (or JSW) worsening that occurred within the same annual follow-up interval as that of IACs use.
      All statistical analyses were performed by using SAS 9.4 (SAS Institute Inc., Cary, NC, USA), and a P value < 0.05 was considered being statistically significant. All tests were two-tailed.

      Results

      Among 2,416 knees (58.9% women, mean age: 62.6 years) eligible for the analysis for the KL grade progression, 153 initiated IACs during their follow-up visits. Of the 153, 148 knees were propensity-score matched to 536 knees that did not receive IACs (Fig. 1). Five knees were excluded from the analysis because no comparison knees could be matched according to propensity score.
      The baseline characteristics of the IACs initiation cohort and comparison cohort are shown in Table I. All baseline covariates were well-balanced between the two cohorts (i.e., standardized difference ≤0.1). During 300 person-years of follow-up, 65 knees in the IACs cohort experienced KL worsening (21.7/100 person-years), whereas 90 knees in the comparison cohort experienced KL worsening during 1,266 person-years of follow-up (7.1/100 person-years). The risk of KL worsening increased more rapidly among the knees in the IACs cohort than those in the comparison cohort (Fig. 2(A), P < 0.001). Compared with no-IACs, the HR of KL worsening for IACs was 3.02 (95% CI, 2.25–4.05) (Table II). The proportional hazard assumption was not violated (P = 0.97). The association did not change materially with propensity-score trimming (HR, 2.95; 95% CI, 2.18–3.99). A total of 62 knees (33 [22.3%] in IACs cohort, 29 [5.4%] in comparison cohort) experienced knee replacements during the follow-up. However, only one knee in IACs cohort had knee replacement during the follow-up period prior to KL worsening assessment. When knee replacement was treated as the competing event (as well as death), the HR of KL worsening for IACs was 2.54 (95% CI, 1.81–3.57).
      Table ICharacteristics in propensity-score matched cohort study of intra-articular corticosteroids (IACs) for KL worsening
      CharacteristicsIACs cohort (n = 148)Comparison cohort (n = 536)Standardized difference
      Sex (female, %)64.266.80.055
      Age (years)64.4 ± 8.664.4 ± 9.10.003
      BMI (kg/m2)30.3 ± 4.130.4 ± 5.00.025
      Race (Whites, %)76.475.40.023
      Education (college or above, %)83.178.90.107
      Physical Activity Level (PASE)146.0 ± 79.5148.2 ± 80.30.027
      Injury (yes, %)36.539.70.067
      WOMAC Knee Pain Score4.5 ± 3.84.5 ± 4.20.001
      NSAIDs Use (yes, %)50.748.00.055
      Baseline KL Grade (KL = 3, %)52.753.90.024
      Charlson Comorbidity Score0.7 ± 1.00.7 ± 1.10.027
      Widespread Pain (yes, %)33.130.80.050
      BMI, body mass index; PASE, Physical Activity Scale for Elderly; WOMAC, The Western Ontario and McMaster Universities Osteoarthritis Index; NSAIDs, non-steroidal anti-inflammatory drugs; KL, Kellgren and Lawrence; IACs, intra-articular corticosteroids.
      Fig. 2
      Fig. 2A: Incidence of KL Worsening in Intra-Articular Corticosteroids Cohort and Comparison Cohort; B: Incidence of JSW Worsening in Intra-Articular Corticosteroids Cohort and Comparison Cohort. KL, Kellgren and Lawrence; JSW, joint space width; IACs, intra-articular corticosteroids.
      Table IIAssociation between intra-articular corticosteroids (IACs) and risk of knee ROA progression
      Knee ROA progressionIACs cohortComparison cohortHazard ratio
      Hazard ratios were adjusted for competing risk.
      95% CI
      Continuous IACs, Hazard ratio 95% CI
      No. knees
      Number of knees with radiographic osteoarthritis progression (total number of knees).
      Incidence rate (1/100 PYs)No. knees
      Number of knees with radiographic osteoarthritis progression (total number of knees).
      Incidence rate (1/100 PYs)
      KL Worsening65 (148)21.790 (536)7.13.02 (2.25, 4.05)4.67 (2.92, 7.47)
      JSW Worsening63 (104)32.099 (388)11.12.92 (2.19, 3.90)3.26 (1.78, 5.96)
      ROA, radiographic osteoarthritis; PYs, person-years; CI, confidence interval; KL, Kellgren and Lawrence; JSW, joint space width; IACs, intra-articular corticosteroids.
      Number of knees with radiographic osteoarthritis progression (total number of knees).
      Hazard ratios were adjusted for competing risk.
      Among 1,533 knees (60.9% women, mean age: 62.5 years) eligible for the analysis for the JSW worsening, 105 initiated IACs during their follow-up visits. Of them, 104 knees were propensity-score matched to 388 knees that did not receive IACs (Fig. 1). One knee was excluded from the analysis because no comparison knee could be identified according to propensity score matching criteria.
      The baseline characteristics of the IACs cohort and its comparison cohort for JSW worsening were well-balanced (eTable 1 in the Supplement). The risk of JSW worsening was higher in the IACs cohort than that in the comparison cohort (Fig. 2(B), P < 0.001). Compared with no-IACs, the HR of JSW worsening for IACs was 2.92 (95% CI, 2.18–3.90) (Table II). The proportional hazard assumption was violated (P < 0.05). The non-proportional HR of JSW worsening for IACs was 2.77 (95% CI, 2.06–3.73). The HR obtained from propensity-score trimming did not change materially (HR, 2.66; 95% CI, 1.96–3.63). A total of 50 knees (32 [30.8%] in IACs cohort, 18 [4.6%] in comparison cohort) had knee replacements during the follow-up. Among them, seven knees had a joint replacement (five occurred in IACs cohort and two in comparison cohort) prior to JSW worsening assessment. The HR of JSW worsening attenuated slightly (HR, 2.36; 95% CI, 1.68–3.32) when knee replacement (as well as death) was treated as a competing event.
      We used marginal structural models to assess the association between continuous IACs and risk of KL grade (or JSW) worsening adjusting for time-fixed and time-varying covariates. Among the 153 knees that initiated IACs use and were eligible for the KL grade worsening, 21 (13.7%) had answered “Yes” to IACs within past 6 months in two annual visits, seven (4.6%) in three visits, and two (1.3%) in four visits, respectively. Of 105 knees that initiated IACs and were eligible for JSW progression, the corresponding number (%) of knees were 15 (14.3%), six (5.7%), and one (1.0%), respectively. A total of 2,416 subjects for KL worsening and 1,533 subjects for JSW worsening were included in the marginal structural models, respectively. For KL worsening, the range of stabilized inverse-probability weights was between 0.1 and 7.7 in IACs group, and between 0.3 and 2.5 in the no-IACs group. For JSW worsening, the range of stabilized inverse-probability weights was between 0.1 and 4.3 in IACs group, and between 0.1 and 4.2 in the no-IACs group. The hazard ratios (HRs) of continuous IACs were 4.67 (95% CI, 2.92–7.47) for KL worsening and 3.26 (95% CI, 1.78–5.96) for JSW worsening, respectively. Treating knee replacement as a censored event attenuated the association of continuous IACs with either KL worsening (HR, 3.42; 95% CI, 2.07–5.63) or JSW worsening (HR, 2.69; 95% CI, 1.50–4.82).
      Our sensitivity analysis indicated that a large residual confounding bias is required to completely explain away the observed association between IACs and the risk of knee ROA progression (eTable 2 and eFigure 1 in the Supplement). For example, a relative risk of an unmeasured confounder(s) with both IACs initiation and risk of KL worsening must be ≥5.49 beyond the confounders that had been adjusted for in the analysis to explain a HR of 3.02 between IACs initiation and the risk of KL worsening observed in the current study. Furthermore, results did not change materially when we additionally added KL grade change (or JSW change) prior to the index visit as a covariate into logistic regression model to calculate the propensity score (HR, 3.00, 95% CI, 1.92–4.68 for KL worsening; HR, 4.09, 95% CI, 2.68–6.23 for JSW worsening). Finally, when KL grade (or JSW) worsening that occurred within the same annual follow-up interval as that of IACs use were excluded from the analysis, initiation of IACs was still associated with an increased risk of KL grade (or JSW) worsening although the effect estimate for JSW worsening was attenuated (Table III).
      Table IIIAssociation between intra-articular corticosteroids (IACs) and risk of knee ROA progression occurring one year after IACs use
      Knee ROA progressionIACs cohortComparison cohortHazard ratio
      Hazard ratios were adjusted for competing risk.
      95% CI
      No. knees
      Number of knees with radiographic osteoarthritis progression (total number of knees).
      Incidence rate (1/100 PYs)No. knees
      Number of knees with radiographic osteoarthritis progression (total number of knees).
      Incidence rate (1/100 PYs)
      KL Worsening30 (93)12.253 (348)5.52.29 (1.52, 3.46)
      JSW Worsening25 (56)16.855 (200)9.51.89 (1.18, 3.05)
      ROA, radiographic osteoarthritis; PYs, person-years; CI, confidence interval; KL, Kellgren and Lawrence; JSW, joint space width; IACs, intra-articular corticosteroids.
      Number of knees with radiographic osteoarthritis progression (total number of knees).
      Hazard ratios were adjusted for competing risk.

      Discussion

      In this cohort study of knee ROA, we found that IACs may be associated with an increased risk of knee ROA progression, and the risk appeared larger with continuous IACs use. These results agree with the recent RCT, which found that repeated IACs led to larger cartilage volume loss, compared with intra-articular saline
      • McAlindon T.E.
      • LaValley M.P.
      • Harvey W.F.
      • Price L.L.
      • Driban J.B.
      • Zhang M.
      • et al.
      Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial.
      . Our findings were independent of the effect of the major confounders, and remained stable in various sensitivity analyses, suggesting that the initiation of IACs and continuous IACs may have a detrimental effect on knee ROA progression.
      To date, two RCTs have evaluated the effect of IACs on knee cartilage
      • Raynauld J.P.
      • Buckland-Wright C.
      • Ward R.
      • Choquette D.
      • Haraoui B.
      • Martel-Pelletier J.
      • et al.
      Safety and efficacy of long-term intraarticular steroid injections in osteoarthritis of the knee: a randomized, double-blind, placebo-controlled trial.
      • McAlindon T.E.
      • LaValley M.P.
      • Harvey W.F.
      • Price L.L.
      • Driban J.B.
      • Zhang M.
      • et al.
      Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial.
      . One study (n = 34 in each arm) reported no difference in mean JSW between IACs and intra-articular saline over 2 years, but the trial may have had insufficient power due to an incorrect parameter estimation when the study was planned
      • Raynauld J.P.
      • Buckland-Wright C.
      • Ward R.
      • Choquette D.
      • Haraoui B.
      • Martel-Pelletier J.
      • et al.
      Safety and efficacy of long-term intraarticular steroid injections in osteoarthritis of the knee: a randomized, double-blind, placebo-controlled trial.
      . Another trial (n = 140), that was adequately powered and utilized more sensitive imaging technique (i.e., magnetic resonance imaging [MRI]), demonstrated a detrimental effect of IACs on cartilage volume, although the amount of cartilage loss was small and may not be clinically meaningful
      • McAlindon T.E.
      • LaValley M.P.
      • Harvey W.F.
      • Price L.L.
      • Driban J.B.
      • Zhang M.
      • et al.
      Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial.
      . However, in both trials, repeated IACs were administered to participants during the study period regardless of the flare of pain or inflammatory signs, which are the usual indications for such injections in clinical practice
      • Luijsterburg P.A.J.
      • Bos P.K.
      • Bierma-Zeinstra S.M.A.
      Long-term intra-articular steroid injections and knee cartilage.
      • Felson D.T.
      Intra-articular corticosteroids and knee osteoarthritis: interpreting different meta-analyses.
      . IACs have been used for pain relief among patients with knee OA for decades. Although IACs might be effective in pain relief for approximately 6 weeks in treating knee OA
      • Juni P.
      • Hari R.
      • Rutjes A.W.
      • Fischer R.
      • Silletta M.G.
      • Reichenbach S.
      • et al.
      Intra-articular corticosteroid for knee osteoarthritis.
      , our findings of the potential detrimental effects of IACs on knee ROA progression raise the concern of its long-term effectiveness on knee pain management, especially previous studies have shown that higher KL grade and decreased JSW were associated with knee pain
      • Neogi T.
      • Felson D.
      • Niu J.
      • Nevitt M.
      • Lewis C.E.
      • Aliabadi P.
      • et al.
      Association between radiographic features of knee osteoarthritis and pain: results from two cohort studies.
      • Oak S.R.
      • Ghodadra A.
      • Winalski C.S.
      • Miniaci A.
      • Jones M.H.
      Radiographic joint space width is correlated with 4-year clinical outcomes in patients with knee osteoarthritis: data from the osteoarthritis initiative.
      . This concern was supported by the findings that participants in the OAI who initiated IACs experienced worsening pain, stiffness, and physical functioning compared with nonusers over 2 years of follow-up
      • Liu S.H.
      • Dube C.E.
      • Eaton C.B.
      • Driban J.B.
      • McAlindon T.E.
      • Lapane K.L.
      Longterm effectiveness of intraarticular injections on patient- reported symptoms in knee osteoarthritis.
      , indicating that IACs did not appear to provide sustained symptom relief for patients with knee OA
      • Juni P.
      • Hari R.
      • Rutjes A.W.
      • Fischer R.
      • Silletta M.G.
      • Reichenbach S.
      • et al.
      Intra-articular corticosteroid for knee osteoarthritis.
      . Nevertheless, since the findings were based on an observational study, we could not rule out the potential residual confounders.
      Biological mechanisms linking corticosteroids to knee ROA progression are not well understood. Given the anti-inflammatory effect of corticosteroids, one would postulate that IACs may reduce the risk of knee ROA progression. However, evidence for inflammation as a predictor for knee OA progression is still inconclusive. Studies reported that MRI-detected synovitis was independently associated with incidence and progression of knee ROA
      • Collins J.E.
      • Losina E.
      • Nevitt M.C.
      • Roemer F.W.
      • Guermazi A.
      • Lynch J.A.
      • et al.
      Semiquantitative imaging biomarkers of knee osteoarthritis progression: data from the foundation for the national Institutes of Health osteoarthritis biomarkers consortium.
      • Felson D.T.
      • Niu J.
      • Neogi T.
      • Goggins J.
      • Nevitt M.C.
      • Roemer F.
      • et al.
      Synovitis and the risk of knee osteoarthritis: the MOST study.
      . Similarly, ultrasound-detected effusion was also a predictor of subsequent joint replacement among patients with knee OA
      • Conaghan P.G.
      • D'Agostino M.A.
      • Le Bars M.
      • Baron G.
      • Schmidely N.
      • Wakefield R.
      • et al.
      Clinical and ultrasonographic predictors of joint replacement for knee osteoarthritis: results from a large, 3-year, prospective EULAR study.
      . In contrast, another study found that MRI-detected synovitis was not associated with cartilage loss in either tibiofemoral or patellofemoral compartment in knee OA
      • Hill C.L.
      • Hunter D.J.
      • Niu J.
      • Clancy M.
      • Guermazi A.
      • Genant H.
      • et al.
      Synovitis detected on magnetic resonance imaging and its relation to pain and cartilage loss in knee osteoarthritis.
      . Furthermore, two systematic reviews have suggested that systemic inflammation measures, such as C-reactive protein, were not associated with knee OA progression
      • Jin X.
      • Beguerie J.R.
      • Zhang W.
      • Blizzard L.
      • Otahal P.
      • Jones G.
      • et al.
      Circulating C reactive protein in osteoarthritis: a systematic review and meta-analysis.
      • Kerkhof H.J.
      • Bierma-Zeinstra S.M.
      • Castano-Betancourt M.C.
      • de Maat M.P.
      • Hofman A.
      • Pols H.A.
      • et al.
      Serum C reactive protein levels and genetic variation in the CRP gene are not associated with the prevalence, incidence or progression of osteoarthritis independent of body mass index.
      . The present study, along with a recent RCT
      • McAlindon T.E.
      • LaValley M.P.
      • Harvey W.F.
      • Price L.L.
      • Driban J.B.
      • Zhang M.
      • et al.
      Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial.
      , reported IACs, especially continuous IACs, are associated with the risk of knee ROA progression. While IACs didn't prevent OA progression it does not preclude that chronic inflammation and synovitis may play a role in the pathogenesis of OA. These results are similar to the findings from previous in vitro and in vivo studies
      • Behrens F.
      • Shepard N.
      • Mitchell N.
      Alterations of rabbit articular cartilage by intra-articular injections of glucocorticoids.
      • Mankin H.J.
      • Conger K.A.
      The acute effects of intra-articular hydrocortisone on articular cartilage in rabbits.
      • Murray R.C.
      • DeBowes R.M.
      • Gaughan E.M.
      • Zhu C.F.
      • Athanasiou K.A.
      The effects of intra-articular methylprednisolone and exercise on the mechanical properties of articular cartilage in the horse.
      • Beitzel K.
      • McCarthy M.B.
      • Cote M.P.
      • Apostolakos J.
      • Russell R.P.
      • Bradley J.
      • et al.
      The effect of ketorolac tromethamine, methylprednisolone, and platelet-rich plasma on human chondrocyte and tenocyte viability.
      • Fahey M.
      • Mitton E.
      • Muth E.
      • Rosenthal A.K.
      Dexamethasone promotes calcium pyrophosphate dihydrate crystal formation by articular chondrocytes.
      • Fubini S.L.
      • Todhunter R.J.
      • Burton-Wurster N.
      • Vernier-Singer M.
      • MacLeod J.N.
      Corticosteroids alter the differentiated phenotype of articular chondrocytes.
      • Miyazaki Y.
      • Tsukazaki T.
      • Hirota Y.
      • Yonekura A.
      • Osaki M.
      • Shindo H.
      • et al.
      Dexamethasone inhibition of TGFβ-induced cell growth and type II collagen mRNA expression through ERK-integrated AP-1 activity in cultured rat articular chondrocytes.
      • Seshadri V.
      • Coyle C.H.
      • Chu C.R.
      Lidocaine potentiates the chondrotoxicity of methylprednisolone.
      • Shen C.
      • Cai G.Q.
      • Peng J.P.
      • Chen X.D.
      Autophagy protects chondrocytes from glucocorticoids-induced apoptosis via ROS/Akt/FOXO3 signaling.
      • Van Offel J.F.
      • Schuerwegh A.J.
      • Bridts C.H.
      • Stevens W.J.
      • De Clerck L.S.
      Effect of bisphosphonates on viability, proliferation, and dexamethasone-induced apoptosis of articular chondrocytes.
      , suggesting that corticosteroids may induce chondrocyte apoptosis
      • Shen C.
      • Cai G.Q.
      • Peng J.P.
      • Chen X.D.
      Autophagy protects chondrocytes from glucocorticoids-induced apoptosis via ROS/Akt/FOXO3 signaling.
      • Van Offel J.F.
      • Schuerwegh A.J.
      • Bridts C.H.
      • Stevens W.J.
      • De Clerck L.S.
      Effect of bisphosphonates on viability, proliferation, and dexamethasone-induced apoptosis of articular chondrocytes.
      , decrease cell viability
      • Beitzel K.
      • McCarthy M.B.
      • Cote M.P.
      • Apostolakos J.
      • Russell R.P.
      • Bradley J.
      • et al.
      The effect of ketorolac tromethamine, methylprednisolone, and platelet-rich plasma on human chondrocyte and tenocyte viability.
      • Seshadri V.
      • Coyle C.H.
      • Chu C.R.
      Lidocaine potentiates the chondrotoxicity of methylprednisolone.
      , suppress the expression of matrix proteins
      • Fubini S.L.
      • Todhunter R.J.
      • Burton-Wurster N.
      • Vernier-Singer M.
      • MacLeod J.N.
      Corticosteroids alter the differentiated phenotype of articular chondrocytes.
      • Miyazaki Y.
      • Tsukazaki T.
      • Hirota Y.
      • Yonekura A.
      • Osaki M.
      • Shindo H.
      • et al.
      Dexamethasone inhibition of TGFβ-induced cell growth and type II collagen mRNA expression through ERK-integrated AP-1 activity in cultured rat articular chondrocytes.
      , or promote calcium pyrophosphate dihydrate crystals formation
      • Fahey M.
      • Mitton E.
      • Muth E.
      • Rosenthal A.K.
      Dexamethasone promotes calcium pyrophosphate dihydrate crystal formation by articular chondrocytes.
      that may accelerate cartilage degeneration.
      The present study has several strengths. The OAI is designed to specifically assess the natural history of knee OA, including its risk factors and clinical management; thus, the findings from our study are likely to represent a real-world setting of clinical practice. Second, we selected patients who initiated IACs during the follow-up period; such a “new user” design should minimize potential selection bias on the study findings. Third, we used two different measurements to assess ROA progression, and the results are similar. Fourth, we performed several sensitivity analyses to assess potential residual confounding and reverse causality bias; and the results did not show a clinically meaningful change, indicating the robustness of our study findings. Finally, the consistency of our results with the recent RCT data
      • McAlindon T.E.
      • LaValley M.P.
      • Harvey W.F.
      • Price L.L.
      • Driban J.B.
      • Zhang M.
      • et al.
      Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial.
      strengthens the validity of potential detrimental effect of IACs on cartilage.
      Several limitations of our study deserve comment. First, information on IACs in the OAI was only assessed within 6 months before each annual visit, but not during the first 6 months between two annual visits, and the number of IACs was not queried. A participant who answered “yes” to IACs in only one visit does not mean he/she only received a single IACs during 12-month period. As a result, the observed HR of knee ROA progression for IACs may have overestimated the detrimental effect of a single IACs as knees categorized as having IACs in only one annual visit may actually have had multiple injections. Second, IACs in the OAI were assessed through a participant's self-report. It is possible that some participants may have mixed IACs with intra-articular hyaluronic acid or may not remember whether they have received IACs within the 6 months before follow-up visit. However, such recall bias, if occurred, would dilute the association between the continuous IACs and the risk of knee ROA progression. Third, our study comparing the risk of knee ROA progression between IACs with no-IACs was susceptible to confounding by indication. Although we used propensity-score matched design (the propensity score has an important balancing property that underlies its value for observational analysis and allows direct estimation of unconfounded risk ratios in cohort studies
      • Glynn R.J.
      • Schneeweiss S.
      • Stürmer T.
      Indications for propensity scores and review of their use in pharmacoepidemiology.
      ) to address this issue, and our quantitative sensitivity analysis
      • VanderWeele T.J.
      • Ding P.
      Sensitivity analysis in observational research: introducing the E-value.
      indicated that the associations of potential residual confounder(s) with either IACs or risk of knee ROA progression must be strong (i.e., relative risk ≥5.0) to completely explain away the observed association, we still can not rule out bias from the residual confounding. Fourth, data on types and concentrations of IACs were not collected in the OAI; thus we were unable to examine the relation of a specific type of IACs and its dose to the risk of knee ROA progression. Finally, in our main analysis, knees that experienced KL grade (or JSW) worsening could have occurred during the same time interval as that of IACs use; thus, reverse causality bias could not be eliminated. However, when we excluded these cases from the analyses, there was still an association between IACs and risk of knee ROA progression.

      Conclusions

      IACs, especially continuous IACs, may be associated with an increased risk of knee ROA progression. However, such findings still warrant replication in other cohorts.

      Contributors

      GL and YZ are joint corresponding authors.
      CZ, JW, GL and YZ had full access to the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. CZ, GL and YZ conceived the study. CZ, GL and YZ were responsible for conception of the study and drafted the manuscript. YZ and HKC were responsible for design of the study. JW, NEL, NL and YZ contributed to preparation and data analysis. NEL, DJH, HKC, TEM and HL contributed to revision of the manuscript. All the authors contributed to the interpretation of the data and critically reviewed the manuscript for publication. All authors read and approved the final manuscript.

      Competing interests

      The authors declare that they have no competing interests.

      Funding

      This work was supported by the National Natural Science Foundation of China ( 81772413 , 81702207 , 81702206 ), the Postdoctoral Science Foundation of Central South University ( 182130 ), the Young Investigator Grant of Xiangya Hospital, Central South University ( 2016Q03 , 2016Q06 ) and the Natural Science Foundation of Hunan Province ( 2017JJ3491 , 2017JJ3492 ). The Osteoarthritis Initiative (OAI) is a public–private partnership comprised of five contracts ( N01-AR-2-2258 ; N01-AR-2-2259 ; N01-AR-2-2260 ; N01-AR-2-2261 ; N01-AR-2-2262 ) funded by the National Institutes of Health , a branch of the Department of Health and Human Services and conducted by the OAI Study Investigators. Private funding partners include Merck Research Laboratories; Novartis Pharmaceuticals Corporation , GlaxoSmithKline and Pfizer . Private sector funding for the OAI is managed by the Foundation for the National Institutes of Health . This manuscript was prepared using an OAI public use dataset and does not necessarily reflect the opinions or views of the OAI investigators, the NIH, or the private funding partners.

      Role of the funding source

      The funding source had no role in the design or conduct of the study; collection, analysis, or interpretation of the data; or writing of the report.

      Ethical approval

      This study received approval from the medical ethical committee, and participants gave written informed consent.

      Acknowledgement

      The contributions of study participants of the Osteoarthritis Initiative are gratefully acknowledged. Everyone who contributed significantly to the work has been listed.

      Appendix A. Supplementary data

      The following is the Supplementary data to this article:

      Reference

        • GBD 2016 DALYs and HALE Collaborators
        Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016.
        Lancet. 2017; 390: 1211-1259
        • Glyn-Jones S.
        • Palmer A.J.
        • Agricola R.
        • Price A.J.
        • Vincent T.L.
        • Weinans H.
        • et al.
        Osteoarthritis.
        Lancet. 2015; 386: 376-387
        • McCabe P.S.
        • Parkes M.J.
        • Maricar N.
        • Hutchinson C.E.
        • Freemont A.
        • O'Neill T.W.
        • et al.
        Brief report: synovial fluid white blood cell count in knee osteoarthritis: association with structural findings and treatment response.
        Arthritis Rheumatol. 2017; 69: 103-107
        • Hall M.
        • Doherty S.
        • Courtney P.
        • Latief K.
        • Zhang W.
        • Doherty M.
        Ultrasound detected synovial change and pain response following intra-articular injection of corticosteroid and a placebo in symptomatic osteoarthritic knees: a pilot study.
        Ann Rheum Dis. 2014; 73: 1590-1591
        • O'Banion M.K.
        • Winn V.D.
        • Young D.A.
        cDNA cloning and functional activity of a glucocorticoid-regulated inflammatory cyclooxygenase.
        Proc Natl Acad Sci U S A. 1992; 89: 4888-4892
        • Rhen T.
        • Cidlowski J.A.
        Antiinflammatory action of glucocorticoids– new mechanisms for old drugs.
        N Engl J Med. 2005; 353: 1711-1723
        • Juni P.
        • Hari R.
        • Rutjes A.W.
        • Fischer R.
        • Silletta M.G.
        • Reichenbach S.
        • et al.
        Intra-articular corticosteroid for knee osteoarthritis.
        Cochrane Database Syst Rev. 2015; : CD005328
        • Bannuru R.R.
        • Schmid C.H.
        • Kent D.M.
        • Vaysbrot E.E.
        • Wong J.B.
        • McAlindon T.E.
        Comparative effectiveness of pharmacologic interventions for knee osteoarthritis: a systematic review and network meta-analysis.
        Ann Intern Med. 2015; 162: 46-54
        • Hochberg M.C.
        • Altman R.D.
        • April K.T.
        • Benkhalti M.
        • Guyatt G.
        • McGowan J.
        • et al.
        American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee.
        Arthritis Care Res (Hoboken). 2012; 64: 465-474
        • McAlindon T.E.
        • Bannuru R.R.
        • Sullivan M.C.
        • Arden N.K.
        • Berenbaum F.
        • Bierma-Zeinstra S.M.
        • et al.
        OARSI guidelines for the non-surgical management of knee osteoarthritis.
        Osteoarthritis Cartilage. 2014; 22: 363-388
        • National Clinical Guideline Centre
        Osteoarthritis: Care and Management in Adults.
        2014 (London)
        • Jevsevar D.S.
        Treatment of osteoarthritis of the knee: evidence-based guideline, 2nd edition.
        J Am Acad Orthop Surg. 2013; 21: 571-576
        • Behrens F.
        • Shepard N.
        • Mitchell N.
        Alterations of rabbit articular cartilage by intra-articular injections of glucocorticoids.
        J Bone Jt Surg Am. 1975; 57: 70-76
        • Mankin H.J.
        • Conger K.A.
        The acute effects of intra-articular hydrocortisone on articular cartilage in rabbits.
        J Bone Jt Surg Am. 1966; 48: 1383-1388
        • Murray R.C.
        • DeBowes R.M.
        • Gaughan E.M.
        • Zhu C.F.
        • Athanasiou K.A.
        The effects of intra-articular methylprednisolone and exercise on the mechanical properties of articular cartilage in the horse.
        Osteoarthritis Cartilage. 1998; 6: 106-114
        • Raynauld J.P.
        • Buckland-Wright C.
        • Ward R.
        • Choquette D.
        • Haraoui B.
        • Martel-Pelletier J.
        • et al.
        Safety and efficacy of long-term intraarticular steroid injections in osteoarthritis of the knee: a randomized, double-blind, placebo-controlled trial.
        Arthritis Rheum. 2003; 48: 370-377
        • McAlindon T.E.
        • LaValley M.P.
        • Harvey W.F.
        • Price L.L.
        • Driban J.B.
        • Zhang M.
        • et al.
        Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial.
        JAMA. 2017; 317: 1967-1975
        • Bliddal H.
        • Henriksen M.
        Osteoarthritis: time to put steroid injections behind us?.
        Nat Rev Rheumatol. 2017; 13: 519-520
        • Duryea J.
        • Neumann G.
        • Niu J.
        • Totterman S.
        • Tamez J.
        • Dabrowski C.
        • et al.
        Comparison of radiographic joint space width with magnetic resonance imaging cartilage morphometry: analysis of longitudinal data from the osteoarthritis initiative.
        Arthritis Care Res (Hoboken). 2010; 62: 932-937
        • Ratzlaff C.
        • Ashbeck E.L.
        • Guermazi A.
        • Roemer F.W.
        • Duryea J.
        • Kwoh C.K.
        A quantitative metric for knee osteoarthritis: reference values of joint space loss.
        Osteoarthritis Cartilage. 2018;
        • Ornetti P.
        • Brandt K.
        • Hellio-Le Graverand M.P.
        • Hochberg M.
        • Hunter D.J.
        • Kloppenburg M.
        • et al.
        OARSI-OMERACT definition of relevant radiological progression in hip/knee osteoarthritis.
        Osteoarthritis Cartilage. 2009; 17: 856-863
        • Choi H.K.
        • Nguyen U.S.
        • Niu J.
        • Danaei G.
        • Zhang Y.
        Selection bias in rheumatic disease research.
        Nat Rev Rheumatol. 2014; 10: 403-412
        • Altman R.D.
        • Gold G.E.
        Atlas of individual radiographic features in osteoarthritis, revised.
        Osteoarthritis Cartilage. 2007; 15: A1-A56
        • Eckstein F.
        • Wirth W.
        • Nevitt M.C.
        Recent advances in osteoarthritis imaging–the osteoarthritis initiative.
        Nat Rev Rheumatol. 2012; 8: 622-630
        • Eckstein F.
        • Collins J.E.
        • Nevitt M.C.
        • Lynch J.A.
        • Kraus V.B.
        • Katz J.N.
        • et al.
        Brief report: cartilage thickness change as an imaging biomarker of knee osteoarthritis progression: data from the foundation for the national Institutes of Health osteoarthritis biomarkers consortium.
        Arthritis Rheumatol. 2015; 67: 3184-3189
        • Ruhdorfer A.
        • Haniel F.
        • Petersohn T.
        • Dorrenberg J.
        • Wirth W.
        • Dannhauer T.
        • et al.
        Between-group differences in infra-patellar fat pad size and signal in symptomatic and radiographic progression of knee osteoarthritis vs non-progressive controls and healthy knees - data from the FNIH biomarkers consortium study and the osteoarthritis initiative.
        Osteoarthritis Cartilage. 2017; 25: 1114-1121
        • Carlesso L.C.
        • Niu J.
        • Segal N.A.
        • Frey-Law L.A.
        • Lewis C.E.
        • Nevitt M.C.
        • et al.
        The effect of widespread pain on knee pain worsening, incident knee osteoarthritis (OA), and incident knee pain: the multicenter OA (MOST) study.
        J Rheumatol. 2017; 44: 493-498
        • Rosenbaum P.R.
        • Rubin D.B.
        The central role of the propensity score in observational studies for causal effects.
        Biometrika. 1983; 70: 41-55
        • Lu N.
        • Misra D.
        • Neogi T.
        • Choi H.K.
        • Zhang Y.
        Total joint arthroplasty and the risk of myocardial infarction: a general population, propensity score-matched cohort study.
        Arthritis Rheumatol. 2015; 67: 2771-2779
        • Seeger J.D.
        • Williams P.L.
        • Walker A.M.
        An application of propensity score matching using claims data.
        Pharmacoepidemiol Drug Saf. 2005; 14: 465-476
        • Austin P.C.
        Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies.
        Pharm Stat. 2011; 10: 150-161
        • Franklin J.M.
        • Rassen J.A.
        • Ackermann D.
        • Bartels D.B.
        • Schneeweiss S.
        Metrics for covariate balance in cohort studies of causal effects.
        Stat Med. 2014; 33: 1685-1699
        • Lin D.Y.
        • Wei L.-J.
        • Ying Z.
        Checking the Cox model with cumulative sums of martingale-based residuals.
        Biometrika. 1993; 80: 557-572
        • Dunkler D.
        • Ploner M.
        • Schemper M.
        • Heinze G.
        Weighted Cox regression using the R package coxphw.
        J Stat Software. 2018; 84: 1-26
        • Hernan M.A.
        • Brumback B.
        • Robins J.M.
        Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men.
        Epidemiology. 2000; 11: 561-570
        • Toh S.
        • Hernandez-Diaz S.
        • Logan R.
        • Rossouw J.E.
        • Hernan M.A.
        Coronary heart disease in postmenopausal recipients of estrogen plus progestin therapy: does the increased risk ever disappear? A randomized trial.
        Ann Intern Med. 2010; 152: 211-217
        • Sturmer T.
        • Rothman K.J.
        • Avorn J.
        • Glynn R.J.
        Treatment effects in the presence of unmeasured confounding: dealing with observations in the tails of the propensity score distribution–a simulation study.
        Am J Epidemiol. 2010; 172: 843-854
        • VanderWeele T.J.
        • Ding P.
        Sensitivity analysis in observational research: introducing the E-value.
        Ann Intern Med. 2017; 167: 268-274
        • Luijsterburg P.A.J.
        • Bos P.K.
        • Bierma-Zeinstra S.M.A.
        Long-term intra-articular steroid injections and knee cartilage.
        JAMA. 2017; 318: 1184
        • Felson D.T.
        Intra-articular corticosteroids and knee osteoarthritis: interpreting different meta-analyses.
        JAMA. 2016; 316: 2607-2608
        • Neogi T.
        • Felson D.
        • Niu J.
        • Nevitt M.
        • Lewis C.E.
        • Aliabadi P.
        • et al.
        Association between radiographic features of knee osteoarthritis and pain: results from two cohort studies.
        BMJ. 2009; 339: b2844
        • Oak S.R.
        • Ghodadra A.
        • Winalski C.S.
        • Miniaci A.
        • Jones M.H.
        Radiographic joint space width is correlated with 4-year clinical outcomes in patients with knee osteoarthritis: data from the osteoarthritis initiative.
        Osteoarthritis Cartilage. 2013; 21: 1185-1190
        • Liu S.H.
        • Dube C.E.
        • Eaton C.B.
        • Driban J.B.
        • McAlindon T.E.
        • Lapane K.L.
        Longterm effectiveness of intraarticular injections on patient- reported symptoms in knee osteoarthritis.
        J Rheumatol. 2018; 45: 1316-1324
        • Collins J.E.
        • Losina E.
        • Nevitt M.C.
        • Roemer F.W.
        • Guermazi A.
        • Lynch J.A.
        • et al.
        Semiquantitative imaging biomarkers of knee osteoarthritis progression: data from the foundation for the national Institutes of Health osteoarthritis biomarkers consortium.
        Arthritis Rheumatol. 2016; 68: 2422-2431
        • Felson D.T.
        • Niu J.
        • Neogi T.
        • Goggins J.
        • Nevitt M.C.
        • Roemer F.
        • et al.
        Synovitis and the risk of knee osteoarthritis: the MOST study.
        Osteoarthritis Cartilage. 2016; 24: 458-464
        • Conaghan P.G.
        • D'Agostino M.A.
        • Le Bars M.
        • Baron G.
        • Schmidely N.
        • Wakefield R.
        • et al.
        Clinical and ultrasonographic predictors of joint replacement for knee osteoarthritis: results from a large, 3-year, prospective EULAR study.
        Ann Rheum Dis. 2010; 69: 644-647
        • Hill C.L.
        • Hunter D.J.
        • Niu J.
        • Clancy M.
        • Guermazi A.
        • Genant H.
        • et al.
        Synovitis detected on magnetic resonance imaging and its relation to pain and cartilage loss in knee osteoarthritis.
        Ann Rheum Dis. 2007; 66: 1599-1603
        • Jin X.
        • Beguerie J.R.
        • Zhang W.
        • Blizzard L.
        • Otahal P.
        • Jones G.
        • et al.
        Circulating C reactive protein in osteoarthritis: a systematic review and meta-analysis.
        Ann Rheum Dis. 2015; 74: 703-710
        • Kerkhof H.J.
        • Bierma-Zeinstra S.M.
        • Castano-Betancourt M.C.
        • de Maat M.P.
        • Hofman A.
        • Pols H.A.
        • et al.
        Serum C reactive protein levels and genetic variation in the CRP gene are not associated with the prevalence, incidence or progression of osteoarthritis independent of body mass index.
        Ann Rheum Dis. 2010; 69: 1976-1982
        • Beitzel K.
        • McCarthy M.B.
        • Cote M.P.
        • Apostolakos J.
        • Russell R.P.
        • Bradley J.
        • et al.
        The effect of ketorolac tromethamine, methylprednisolone, and platelet-rich plasma on human chondrocyte and tenocyte viability.
        Arthroscopy. 2013; 29: 1164-1174
        • Fahey M.
        • Mitton E.
        • Muth E.
        • Rosenthal A.K.
        Dexamethasone promotes calcium pyrophosphate dihydrate crystal formation by articular chondrocytes.
        J Rheumatol. 2009; 36: 163-169
        • Fubini S.L.
        • Todhunter R.J.
        • Burton-Wurster N.
        • Vernier-Singer M.
        • MacLeod J.N.
        Corticosteroids alter the differentiated phenotype of articular chondrocytes.
        J Orthop Res. 2001; 19: 688-695
        • Miyazaki Y.
        • Tsukazaki T.
        • Hirota Y.
        • Yonekura A.
        • Osaki M.
        • Shindo H.
        • et al.
        Dexamethasone inhibition of TGFβ-induced cell growth and type II collagen mRNA expression through ERK-integrated AP-1 activity in cultured rat articular chondrocytes.
        Osteoarthritis Cartilage. 2000; 8: 378-385
        • Seshadri V.
        • Coyle C.H.
        • Chu C.R.
        Lidocaine potentiates the chondrotoxicity of methylprednisolone.
        Arthroscopy. 2009; 25: 337-347
        • Shen C.
        • Cai G.Q.
        • Peng J.P.
        • Chen X.D.
        Autophagy protects chondrocytes from glucocorticoids-induced apoptosis via ROS/Akt/FOXO3 signaling.
        Osteoarthritis Cartilage. 2015; 23: 2279-2287
        • Van Offel J.F.
        • Schuerwegh A.J.
        • Bridts C.H.
        • Stevens W.J.
        • De Clerck L.S.
        Effect of bisphosphonates on viability, proliferation, and dexamethasone-induced apoptosis of articular chondrocytes.
        Ann Rheum Dis. 2002; 61: 925-928
        • Glynn R.J.
        • Schneeweiss S.
        • Stürmer T.
        Indications for propensity scores and review of their use in pharmacoepidemiology.
        Basic Clin Pharmacol Toxicol. 2006; 98: 253-259