Comparative Effectiveness of Hyaluronic Acid Injections for Management of Knee Osteoarthritis

Osteoarthritis (OA) is a degenerative joint disease in which there is an imbalance between the breakdown and repair of the joint tissue.Intraarticular hyaluronic acid (HA) injections are used for the management of knee OA. The pooled estimates of randomized clinical trials in various meta-analyses do not have consistent conclusions, some conclude no benefit while others conclude small benefit (reducing the pain or improving the daily functioning) to overall benefit (alleviating pain as well as improving daily functioning).The variation in conclusions is due to the heterogeneity in the methodology of the clinical trials included in these meta-analyses as well as a difference in the interpretation of clinical findings.Furthermore, most of the trials (63%) evaluating efficacy of HA are industry funded, raising the concern for publication bias. Furthermore Although the evidence favoring the use of HA injections is limited, these injections still have a significant market share with an annual sale of $725 million per year. Given the high dollar amount spent on the intervention, evaluation of its effectiveness in real world setting is important. Currently, there is limited and inconclusive evidence supporting use of HA injections for management of knee OA. The primary objective of the current study is to evaluate the effectiveness of HA injections in the management of knee OA. Specifically, the investigators will compare the risk of any surgical intervention of knee as a primary outcome among the knee OA patients who are exposed to HA injections with those who are not exposed to HA injections (HA non-users) and those who are exposed to intra-articular corticosteroid (CS) injections. Three separate outcome definitions, which includes: i) composite surgical outcome measure (includes total knee replacement, partial knee replacement, arthroscopic procedures, osteotomy and free-floating inter-positional devices), ii) total and partial knee replacements only and iii) total knee replacement (TKR) only will be used to compare the effectiveness of HA users with HA non-users and CS users.

This study will be conducted using knee OA patients aged 40 years and above from the Lifelink Plus claims data (2006-2015). Knee OA patients with a specialist visit and a recent history of medications used for pain management (proxy for moderate-severe pain) will be identified. Patients will be classified into 3 groups: 1) HA users, 2) CS users and 3) HA non-users based on exposure/non-exposure to these interventions within first 90 days after specialist visit. HA users will be matched separately with two comparison (CS users and HA non-users) groups using a propensity score matching approach to reduce the imbalance between the intervention and comparison group. HA users will be matched separately to each comparison group (1:1 ratio) using a greedy matching approach within a predefined caliper (0.2 of the pooled standard deviation of the logit of the PS). Cox models will be used to compare the risk of 1) any surgical interventions for knee, 2) TKR and partial knee replacements only and 3) TKR only, among HA users and the comparison groups. For both comparisons, investigators will use four approaches to compare the risk of each outcome measure: 1) Unadjusted bivariate analysis, 2) Adjusted multivariate analysis, 3) Propensity score-matched sample and 4) Inverse probability weighting (IPW) using propensity score.

Inclusion criteria

• Patients with a specialist visit (orthopedic surgeon, physical medicine and rehabilitation, orthopedic, physical therapy, and rheumatologist) will enter in cohort and the date of the first visit will be considered as index date.
• Patients should have at least one claim with a primary diagnosis of knee osteoarthritis (ICD-9-Center for Medicare (CM) code 715.x6) on the index date
• Patients with a specialist visit should have at least one claim for pain medications (NSAIDs or opioids) within 90 days prior to or on the day of visit.
• The age of a patient should be at least 40 years on the date of their index date
• Patients to have continuous eligibility for pharmacy and medical benefits for at least six months prior to their index diagnosis.
• Patients to have at least three months of continuous eligibility for pharmacy and medical benefits after their index diagnosis.

Exclusion criteria

• Patients with a procedure claim for HA or CS in the pre-index period
• Patients with a procedure claim for knee OA surgical procedures (arthroscopic procedures, osteotomy, free-floating interpositional devices, partial and total knee replacement) in the pre-index period.
• In order to focus on knee OA pain, patients with a claim for joint fusion, rheumatoid arthritis, knee fracture, post-traumatic arthritis, avascular necrosis, benign/malignant bone tumors and Paget's disease in the pre-index period will be excluded