Osteoarthritis Year in Review 2015: Clinical

Abstract

The purpose of this review is to highlight clinical research in osteoarthritis. A literature search was conducted using PubMed (http://www.ncbi.nlm.nih.gov/pubmed/) with the search terms “osteoarthritis [All Fields] AND treatment [All Fields]” and the following limits activated: humans, English language, all adult 19+ years, published between April 1, 2014 and April 1, 2015. A second literature search was then conducted with the search terms “osteoarthritis [All Fields] AND epidemiology [All Fields]”, with the same limits. Reports of surgical outcome, case series, surgical technique, tissue sample or culture studies, trial protocols, and pilot studies were excluded. Of 1523, 148 were considered relevant. Among epidemiologic and observational clinical studies, themes included physical activity, early knee OA, and confidence/instability/falls. Symptom outcomes of pharmacologic treatments were reported for methotrexate, adalimumab, anti-nerve growth factor monoclonal antibodies, strontium ranelate, bisphosphonates, glucosamine, and chondroitin sulfate, and structural outcomes of pharmacologic treatments for strontium ranelate, recombinant human fibroblast growth factor 18, and glucosamine and chondroitin sulfate. Symptom outcomes of non-pharmacologic interventions were reported for: neuromuscular exercise, quadriceps strengthening, weight reduction and maintenance, TENS, therapeutic ultrasound, stepped care strategies, cognitive behavior therapy for sleep disturbance, acupuncture, gait modification, booster physical therapy, a web-based therapeutic exercise resource center for knee OA; hip physical therapy for hip OA; and joint protection and hand exercises for hand OA. Structure outcomes of non-pharmacologic interventions were reported for patellofemoral bracing.

INTRODUCTION

This is an important era in the clinical investigation of osteoarthritis (OA); a confluence of advances positions investigators to accelerate the pace of work and gains in knowledge. Towards this mission, an annual review can help to organize the large amount of work accomplished in the past year, a particularly useful task for a field as broad and heterogeneous as clinical OA. While some themes emerged, a number of studies did not fall into a theme per se. The OARSI definition of OA is: “a disorder involving movable joints characterized by cell stress and extracellular matrix degradation initiated by micro- and macro-injury that activates maladaptive repair responses including pro-inflammatory pathways of innate immunity. The disease manifests first as a molecular derangement (abnormal joint tissue metabolism) followed by anatomic, and/or physiologic derangements (characterized by cartilage degradation, bone remodeling, osteophyte formation, joint inflammation and loss of normal joint function), that can culminate in illness” (http://oarsi.org/research/standardization-osteoarthritis-definitions).

Among the most important concepts that emerged in this annual review of the literature is continued and further appreciation of stage of pre-disease or disease and clinical phenotype, both in terms of prognosis and suitability and likelihood of success of specific interventions. The ultimate goal of intervention that delays disease and disability progression in this heterogeneous condition will necessitate clear delineation of the phenotypes and disease stages that are encompassed by the label of OA. In recognition of this, OARSI has made the following call:

New specific and sensitive disease endpoints are critically needed to alleviate roadblocks to development of disease modifying therapeutics for OA. A key step in this process is the development of standardized definitions of OA. Standardization of OA definitions would aid communication across the field and help advance drug development for OA and research by achieving consensus on globally recognized definitions of disease and globally recognized standards for classifying the disease. We anticipate that these definitions could facilitate communication about the disease among industry and non-industry researchers, regulatory agencies, funding agencies, third party payers, and patients.

We further anticipate that these definitions would be maintained by OARSI and be subjected to regular refinement as new scientific advances demand. Definitions proposed are not intended to distinguish an OA patient uniquely from patients with other forms of arthritis; but rather, they are intended to provide definitions of the disease process that supersede the many and varied OA phenotypes, to spur scientific advances, and facilitate communication with regulatory agencies. The draft definitions can be viewed as the building blocks for defining OA phenotypes. We fully acknowledge that these building blocks are likely most applicable to knee and hip OA, possibly helpful for hand OA, but will require modification for spine OA. (http://oarsi.org/research/standardization-osteoarthritis-definitions)

The purpose of the current review is to highlight clinical research in OA, particularly in the realms of epidemiology, observational clinical studies, pharmacologic treatment, and nonpharmacologic interventions and strategies. Although an attempt to summarize clinical research in OA in the past year is valuable, it must be emphasized that this is a narrative review and that it is not feasible to summarize all of the important findings of each of these papers.

METHODS

A literature search was conducted using PubMed (http://www.ncbi.nlm.nih.gov/pubmed/) with the search terms “osteoarthritis [All Fields] AND treatment [All Fields]” and the following limits activated: humans, English language, all adult 19+ years, published between April 1, 2014 and April 1, 2015. A second literature search was then conducted with the search terms “osteoarthritis [All Fields] AND epidemiology [All Fields]”, with the same limits. These searches were repeated using Embase (http://www.embase.com/). Reports of surgical outcome, case series, surgical technique, tissue sample or culture studies, trial protocols, pilot studies and meeting abstracts were excluded as were reports focusing on imaging, biomarkers, and rehabilitation, as these topics are being summarized within other reviews in this issue. Of 1523 papers identified, 148 were considered relevant

RESULTS

Papers are described under the heading (below) which corresponds with the primary goal of the reported work. However, some papers could have been included in more than one category. Table 1 lists all papers falling within a given category.

TABLE 1.

Osteoarthritis Papers Identified within Each Category

Category of research	Citations identified falling within category
Physical activity and OA	(1-9)
Early OA	(10-15)
Confidence/instability/falls and OA	(16-20)
Cross-sectional studies dealing with prevalent knee OA	(21-31)
Longitudinal studies dealing with incident knee OA	(3, 12, 13, 32-36)
Longitudinal studies dealing with knee OA disease progression	(37-43)
Physical functioning in knee OA – cross-sectional studies	(5, 17, 19, 44-49, 58, 64, 68)
Physical functioning in knee OA – longitudinal studies	(6, 14, 50-53)
Disability	(7)
Hip OA	(54-58)
Hand OA	(59-65)
Foot OA	(66-69)
Pain/symptoms and OA	(2, 4, 10, 11, 13, 14, 16, 21, 25, 28, 29, 38, 40, 45, 58, 61, 64, 69-85)
Symptom outcome of pharmacological treatments for OA	(86-101)
Structural outcome of pharmacological treatments for OA	(102-106)
Symptom outcome of non-pharmacological treatments for OA	(107-129)
Structural outcome of non-pharmacological treatments for OA	(132)
Prevalence and impact of OA	(133-140)
Health services research in OA	(9, 141-147)
Research priorities in OA	(148, 149)

Epidemiologic and Observational Clinical Studies

Manuscripts are organized into certain themes that emerged (physical activity, early OA, confidence/instability/falls), followed by sections devoted to cross-sectional studies of prevalent knee OA, longitudinal studies dealing with incident knee OA, longitudinal studies dealing with knee OA progression, cross-sectional studies of physical functioning in knee OA, longitudinal studies of physical functioning in knee OA, hip OA, hand OA, foot OA, and pain.

Physical Activity

Papers dealing with physical activity reported on the association of activity with incident knee OA and with joint pain and stiffness, factors associated with activity avoidance, impact of sedentary activity, effect of walking on incident function limitation, qualitative analysis of symptom impact on activity, impact of time in light intensity activity on disability outcomes, and association with health-related utility.

In middle-aged women in the Australian Longitudinal Study on Women's Health, physical activity between 47 and 58 years of age was associated with lower risk of joint pain and stiffness nine years later (1). A qualitative study in Canada characterized consequences of symptoms on physical activity in persons with self-reported knee OA or symptoms (2). Meeting physical activity guidelines was not associated with incident radiographic or symptomatic knee OA in middle-aged or older adults in the Johnston County Osteoarthritis Project (3). Knee pain and lower vitality were associated with activity avoidance (4).

Being less sedentary was associated with better function in the Osteoarthritis Initiative, independent of moderate-vigorous physical activity minutes (5). More walking was associated with a lower risk of incident function limitation in persons with or at higher risk for knee OA in the Multicenter Osteoarthritis Study (6). Greater daily time in light intensity physical activity was associated with reduced onset and progression of disability in Osteoarthritis Initiative participants (7). There was a graded association between sedentary behavior and elevated systolic blood pressure, independent of moderate-vigorous physical activity minutes, in persons with or at higher risk for knee OA in the Osteoarthritis Initiative (8). In another report utilizing the Osteoarthritis Initiative, physical activity level was associated with health-related utility (9).

Early OA

Papers dealing with early OA reported on the first activities to become painful in knee OA, prediction of early knee OA, significance of minor radiographic features, significance of preradiographic MRI lesions, and impact of rapid radiographic change in early OA.

In persons without radiographic OA in the Osteoarthritis Initiative, incident radiographic OA was associated with a prior trajectory of increasing knee pain, stiffness, and difficulties with functional tasks; over the 4-year study period there was no change in symptoms in the control knees which did not develop OA (10). Pain and difficulty with activities associated with higher dynamic loading were associated with longer prodromal phases (10). In persons with or at higher risk for knee OA in the Osteoarthritis Initiative, knee pain was most likely to first appear during weightbearing activities involving bending, such as using stairs (11). Questionnaire variables, genetic markers, other-site OA, and biochemical markers added only modestly to prediction of knee OA incidence using age, gender, and body mass index (BMI) in the Rotterdam Study-I; doubtful minor x-ray features strongly predicted future knee OA (12). In persons Kellgren and Lawrence radiographic grade 0 in both knees in the Osteoarthritis Initiative, baseline cartilage damage, bone marrow lesions, and meniscal tears were associated with persistent symptoms in the next 4 years (13). In persons with early symptomatic knee OA in the Osteoarthritis Initiative and in the Cohort Hip and Cohort Knee Study, rapid radiological change was associated with worsening of pain and function (14). Young and active athletes in a study in Qatar had a greater risk of manifesting early OA features (15).

Confidence/Instability/Falls

The spectrum of knee confidence, instability, and falls emerged as a theme, including manuscripts dealing with factors associated with knee confidence, consequences of knee buckling, fall risk in the setting of OA, the bone mineral density/fracture relationship in OA, and the combined impact of falls and OA on function.

Worse knee confidence was associated with pain, self-reported knee instability, weakness, and varus-valgus motion during gait in the setting of medial tibiofemoral OA in a study in Australia (16). Knee buckling and the sensation of instability without buckling was associated with fear of falling, poor balance confidence, activity limitations, and poor function in the Multicenter Osteoarthritis Study (17). In a longitudinal study with mean follow-up of 6 years from the Johnston County Osteoarthritis Project, the odds of falling increased with an increasing number of lower limb symptomatic OA joints (from 53% higher odds with 1 joint to 85% higher odds with 3-4 OA joints) (18). Fall history and knee OA (vs. neither or only one) was associated with worse KOOS-QoL and SF12 Physical Component Summary scores, in persons with or at higher risk for knee OA in the Osteoarthritis Initiative (19). Fracture risk was higher in women with OA (Dubbo Osteoporosis Epidemiology Study) than those without OA; however, the association was mainly observed in women with osteopenic BMD (for vertebral and limb fracture, but not hip fracture) and normal BMD and not in those with osteoporosis (20).

Cross-sectional Studies Dealing with Prevalent Knee OA

Type 3 finger length pattern, an indicator of prenatal androgen exposure, was associated with symptomatic (but not radiographic) knee OA, chronic pain, and hand OA in the Rotterdam Study (21). Almost half of the association between elevated BMI and knee OA was explained by serum leptin level in the Maintenance of Balance, Independent Living, Intellect, and Zest in the Elderly Boston Study (22). In the Netherlands Epidemiology of Obesity study, knee OA was associated with weight and fat-free mass, adjusting for metabolic factors, and hand OA was associated with the metabolic syndrome, adjusting for weight (23). In a study of patients undergoing total knee or hip replacement due to primary OA in Newfoundland, Canada, metabolomics analysis of synovial fluid samples demonstrated metabolically distinct subgroups (24).

Knee extensor and flexor muscle strength were associated with knee symptoms but not radiographic severity of knee OA in the Osteoarthritis Initiative (25). In the Multicenter Osteoarthritis Study, as compared to limbs without knee OA, limbs with lateral knee OA had a reduced femoral offset, an increased height of hip center, a more valgus neck-shaft angle, and an increased abductor angle, and limbs with medial knee OA had a more varus neck-shaft angle and a decreased abductor angle (26). The relationship between the external knee adduction moment and directly measured medial tibiofemoral contact forces was generally good but varied across subjects and activities in a study in the U.S., illustrating the limitations of relying solely on the adduction moment to infer medial joint loading when excessive directed cocontraction of muscles exists (27).

Patellofemoral OA was common in middle aged individuals with chronic patellofemoral pain in a study in Australia (28). In persons with predominant patellofemoral OA in a U.K. study, greater arthrogenous quadriceps muscle inhibition was associated with more severe knee pain but less severe bone marrow lesion score (29).

In a study including cases from the U.K. High Bone Mass study vs. unaffected family members and general population controls from the Chingford and Hertfordshire cohort studies, interrelationships were observed between osteophytes, enthesophytes, and high bone mass, supporting a subset of knee OA characterized by excess bone formation (30). While overall number and size of bone spurs were similar in patients with psoriatic arthritis and hand OA in a study in Germany utilizing high-resolution peripheral quantitative computer tomography, anatomic sites differed: in psoriatic arthritis, radial side; and in OA, palmar and dorsal quadrants (31).

Longitudinal Studies Dealing with Incident Knee OA

Effusion-synovitis and Hoffa-synovitis were associated with incident radiographic OA in Osteoarthritis Initiative participants (32). In another report from the Osteoarthritis Initiative, among knees without knee OA at baseline, recent knee injury was associated with accelerated development of end-stage knee OA (33). In middle aged, overweight women without knee OA in the Prevention of Knee Osteoarthritis in Overweight Females study, varus alignment was associated with incident radiographic OA; the association for valgus was borderline (34). Serum leptin at baseline was associated with MRI lesions 10 yrs later in the Michigan Study of Women's Health Across the Nation (35). A systematic review and meta-analysis showed that knee extensor muscle weakness was associated with an increased risk of developing symptomatic knee OA (36).

Longitudinal Studies Dealing with Knee OA Disease Progression

Frequent milk consumption was associated with reduced progression of knee OA in women in the Osteoarthritis Initiative (Lu). Vitamin D deficiency at 30 or 36 months was associated with 24-48 month knee OA progression in Osteoarthritis Initiative participants (37). In older adults with and without OA in the Tasmanian Older Adult Cohort study, moderate vitamin D deficiency was associated with incident or worsening knee pain over 5 years (38). In the Health, Aging and Body Composition Study, very low plasma phylloquinone (vitamin K1) was associated with progression of articular cartilage and meniscal damage in older individuals with and without OA (39).

In knees with medial meniscal tears in a study in Australia, weight gain was associated with concurrent cartilage loss and pain (40). In a 1-year ultrasound study of knee OA in The Netherlands, effusion and/or synovial hypertrophy were present in >40% at some time in the year and fluctuated; Baker's cyst and meniscal extrusion were present and stable (41). High total body BMD was associated with an increase in femoral cartilage thickness, high spine BMD with an increase in femoral and lateral tibial cartilage thickness, and high medial tibial subchondral BMD with an increase in medial tibial cartilage thickness in persons with knee OA in the Tasmanian Older Adult Cohort study (42). In the Osteoarthritis Initiative, a larger medial subchondral surface ratio (a measure of compartment-specific size mismatch) was associated with a reduced risk of incident knee symptoms; a larger lateral ratio was associated with a reduced risk of incident and progressive lateral OA (43).

Physical Functioning in Knee OA – Cross-sectional Studies

Knee OA was associated with a worse SF-36 Physical Component Summary score, as were obesity, comorbidities, and low fat free mass in The Netherlands Epidemiology of Obesity study (44). A knee-specific function measure was associated with pain status in the contralateral knee in Osteoarthritis Initiative participants; the chair stand performance measure best distinguished knee pain groups (45). Small increments of muscle strength were associated with clinically relevant differences in a knee-specific function measure in the Osteoarthritis Initiative (46)

Individuals with symptomatic knee OA in a study in the U.S. had lower energetic recovery at self-selected walking speeds vs. healthy controls; findings suggested reduced effective exchange of potential and kinetic energy and increased muscular work required to control center of mass movements (47). In older adults with advanced knee OA and chronic pain in a U.S. study, age, function self-efficacy, and opioid use were associated with slower gait speed (48). In a large dataset comprising cohorts from 6 European countries, lower extremity OA at the hip or knee was associated with worse physical performance (walking speed, chair rises, balance); those at highest risk had OA at both hip and knee (49).

Physical Functioning in Knee OA – Longitudinal Studies

In the Cohort Hip and Cohort Knee study, 3-year decrease in muscle strength was associated with concurrent worsening in performance based measures of function (50). In another study of this cohort, 3 trajectories of function outcome were identified; moderate and poor outcome groups were characterized by younger age, higher BMI, greater pain, bony tenderness, reduced knee flexion, hip pain, osteophytosis, more comorbidities, and lower vitality (51). In a cohort of Chinese community-living older adults, OA was among the factors associated with worsening in frailty status (52). In persons with knee OA, baseline sleep disturbance was associated with greater increase in depression and function limitation (53).

Hip OA

In the Johnston County Osteoarthritis Project, variations in proximal femur shape by active shape modeling were modestly associated with incident radiographic hip OA (54). Cam-type femoroacetabular impingement and mild acetabular dysplasia were each associated with development of radiographic hip OA and total hip replacement in the Chingford 1000 Women Study (55). In the Danish working population, cumulative lifting was modestly associated with total hip replacement in men, as was high BMI at age 25 years and gain in BMI in both men and women (56). In the Australian Diabetes, Obesity and Lifestyle Study, low birth weight and preterm birth were associated with total hip replacement in adult life (57). In patients scheduled to undergo total hip replacement in a U.S. study, hip pain was associated with self-report and performance measures of function, while hip abduction strength was associated only with performance measures (58).

Hand OA

There were several papers dealing with hand OA, including some dealing with the association between MRI-based tissue lesions and progression and erosion development, inflammatory features and hand OA progression, hand OA and coronary heart disease events, correlates of pain in hand OA, erosive OA as a more severe form of disease, and prevalence and burden of carpometacarpal (CMC) OA.

In a study of the Oslo hand OA Cohort, MRI bone marrow lesions and synovitis were associated with radiographic joint space narrowing; synovitis, bone marrow lesions, osteophytes, and malalignment were associated with development of radiographic erosions (59). Ultrasound inflammatory features, especially when persistent, were associated with radiographic progression in clinic patients in The Netherlands who met ACR criteria for hand OA (60). In persons with radiographic hand OA in the Musculoskeletal Pain in Ullensaker Study, diabetes, lower education, female gender, familial OA, widespread pain, poor mental health, more finger joints with ultrasound-detected synovitis and radiographic OA were associated with hand pain (61). Symptomatic (but not radiographic) hand OA was associated with increased risk of coronary heart disease events in the Framingham Heart Study (62).

In 2 population-based studies in North Staffordshire, a similar frequency of joint involvement and patterning was found in erosive OA and severe non-erosive OA, suggesting that erosive OA may be a more severe form of hand OA; erosive OA (vs. non-erosive) was associated with metabolic syndrome, especially dyslipidemia (63). First CMC erosive disease was found in 2.2% of the population-based North Staffordshire cohort; 0.5% had combined 1^st CMC and interphalangeal (IP) erosive disease (64). First CMC erosive disease was associated with greater pain but similar function (vs. non-erosive OA) (64). The prevalence of doctor-diagnosed 1^st CMC OA was 1.4% (2.2% in women and 0.6% in men) in a large health care database in Sweden (65).

Foot OA

Studies relating to foot OA focused on factors associated with hallux valgus, factors associated with 1^st metatarsophalangeal (MTP) OA severity, prevalence of foot OA, and impact of foot pain on function.

In the Johnston County Osteoarthritis Project, hallux valgus was associated with female gender, African-American race, older age, pes planus, knee/hip OA, and inversely with higher BMI (66). In the Clinical Assessment Study of the Foot in the U.K., first MTP OA severity was associated with dorsal hallux and 1^st MTP pain, hallux valgus, 1^st IP hyperextension, and dorsal hallux and 1^st MTP keratotic lesions, and decreased 1^st MTP dorsiflexion, ankle/subtalar eversion and ankle dorsiflexion range (67). In another report from this study, population prevalence of symptomatic radiographic foot OA was reported to be 16.7%: at the 1^st MTP in 7.8%, at the 1^st cuneometatarsal joint in 3.9%, at the 2^nd cuneometatarsal in 6.8%, at the navicular first cuneiform in 5.2%, and at the talonavicular in 5.8% (68). At most sites, prevalence was greater in women, increased with age, and was higher with worse socioeconomic status. Seventy-five% with symptomatic radiographic foot OA reported disabling symptoms (68). In persons with symptomatic knee OA in the Osteoarthritis Initiative, 25% had foot pain (more often bilateral), the presence of which was associated with worse function (69).

Pain

Papers focusing on pain in OA frequently dealt with mechanisms, trajectories, and risk profiles.

In a cross-sectional study using the fifth Korean National Health and Nutrition Examination Survey, a metabolic syndrome/knee OA association was largely explained by excess weight and not insulin resistance; accumulation of metabolic syndrome components was associated with pain severity (70). In persons recruited from the community and clinic with symptomatic knee OA by ACR criteria, different quantitative sensory testing measures were associated with clinical pain in African-Americans and non-Hispanic Caucasians; reduced pain inhibition was important in all groups (71). Women in the Multicenter Osteoarthritis Study reported greater knee pain than men regardless of KL grade; differences decreased with adjustment for widespread pain (72). Higher BMI was associated with greater knee pain accounting for OA severity in persons with or at high risk for knee OA in the Osteoarthritis Initiative (73).

In the Generation of Models, Mechanism & Markers for Stratification of Osteoarthritis Patients study, 3 patterns of synovitis were identified in knee OA: 1) mainly patellar sites, associated with KOOS pain and ICOAP constant pain; 2) mostly near the ACL, not associated with pain or disease severity; 3) mainly at loose bodies, associated with disease severity (74). In another study of patients with knee OA by ACR clinical criteria, there was no association between ultrasound features and knee pain severity by KOOS pain or numeric rating scales (75).

Papers dealing with sensitivity included a report that pressure-pain threshold was associated with single and multijoint symptoms but not asymptomatic knee or hip OA in the Johnston County Osteoarthritis Project (76). A higher neuropathic pain questionnaire score was associated with widespread reduction in the pressure-pain threshold in persons with symptomatic, radiographic knee OA in the U.K. (77). Pressure-pain threshold and temporal summation were associated with pain severity but not OA presence, severity, or duration in persons with or at higher risk for knee OA in the Multicenter Osteoarthritis Study (78).

Other papers characterized pain experience and trajectories in OA. Pain flares in persons with symptomatic knee OA were described most often by quality (sharp), timing (seconds-minutes), antecedents, and consequences (79). In the Mechanical Factors in Arthritis of the Knee study in the U.S., a pain outcome measure was derived based on development of unpredictable pain; good 2-year outcome was less likely with greater catastrophizing and more likely with greater self-efficacy (80). Five WOMAC-pain trajectories over 6 years were identified in persons with symptomatic knee OA in the Osteoarthritis Initiative, none with substantial worsening; greater pain was associated with worse KL grade, obesity, depression, comorbidity, gender (female), non-Caucasian race, lower education, and younger age (81). Mild/non-progressive, progressive, moderate, improving, and severe/non-improving pain trajectories were identified in persons with symptomatic OA in the Knee Clinical Assessment Study; in Osteoarthritis Initiative participants, the progressive and improving groups were less evident (82).

Among other studies dealing with pain, nocturnal knee pain and worse sleep quality were associated with knee OA severity in participants from the Iwaki Health Promotion Project (83). More people with high knee pain/low disease burden of knee OA (vs. high pain/high OA, low pain/high OA, low pain/low OA) had widespread pain, in persons with or at higher risk for knee OA in the Multicenter Osteoarthritis Study (84). In a placebo-controlled, double-blind trial in participants with first CMC OA of the right hand, after naproxen (vs. placebo), fMRI activity was reduced in brain regions commonly associated with pain perception; changes in perceived pain intensity were associated with changes in activity in regions previously associated with pain intensity (85).

Symptom Outcome of Pharmacologic Treatments

Symptom outcomes of pharmacologic treatments were reported for methotrexate, adalimumab, anti-nerve growth factor monoclonal antibodies, strontium ranelate, bisphosphonates, glucosamine, and chondroitin sulfate. Reviews included a Cochrane review of topical rubefacients containing salicylates for chronic conditions and systematic review and network meta-analysis to analyze comparative effectiveness of pharmacologic interventions for knee OA.

In persons with symptomatic knee OA in a double-blind randomized placebo-controlled trial in Egypt, methotrexate reduced pain and improved function at 28 week follow-up (86). In a randomized, double-blind, parallel group, placebo-controlled multicenter trial in France, adalimumab was not superior to placebo to alleviate pain in hand OA not responsive to analgesics and NSAIDs (87). In persons with symptomatic knee OA by ACR clinical criteria, strontium ranelate over 3 years was associated with knee symptom and function improvement in the Strontium Ranelate Efficacy in Knee Osteoarthritis Trial (88). Bisphosphonate users with symptomatic knee OA in the Osteoarthritis Initiative had lower numeric rating scale pain scores vs. non-users (until year 4), but did not differ in WOMAC pain or function (89).

Studies of anti-nerve growth factor monoclonal antibodies included a report that individuals with knee or hip OA by ACR criteria with radiographic confirmation receiving partial symptom relief with NSAIDs may receive greater benefit with tanezumab monotherapy; adverse event frequency was higher with tanezumab than with NSAIDS, and was highest with the combination (90). Adding tanezumab to diclofenac SR improved pain, function, global assessment in patients with knee or hip OA, but higher adverse events in the combination group suggests this approach is unfavorable (91). Fasinumab, in a double-blind, placebo-controlled, parallel group, exploratory study, was generally well tolerated; all 3 doses were better than placebo for walking knee pain and WOMAC (92).

A meta-analysis of placebo-controlled trials of glucosamine for primary or secondary OA in any synovial joint revealed that most heterogeneity was explained by brand; trials using one product had better outcome but large inconsistency, and low risk of bias trials showed small effect sizes (93). In a double-blind randomized clinical trial, non-inferiority trial vs. celecoxib, combined chondroitin sulfate and glucosamine had comparable efficacy to celecoxib for symptoms, function, and joint swelling/effusion in persons with symptomatic knee OA at 6 months (94).

Studies of NSAIDS include a report that long-term NSAID use in the Osteoarthritis Initiative was associated with modest clinical (but not significant) change in stiffness, function, and structure change at the knee (95). Fixed-dose tramadol-diclofenac resulted in greater pain reduction and was well tolerated vs. tramadol-paracetamol for acute OA flare (96). There was no association between traditional NSAIDs and nonfatal acute myocardial infarction when background cardiovascular risk was low-intermediate; there was a moderate association in those at high risk or with NSAID use >365 days, in a nested case-control study using a primary care database in Spain (97). Celecoxib was as effective and safe as naproxen for knee OA in a sample of Hispanic individuals (98).

A Cochrane review concluded that evidence does not support topical rubefacients containing salicylates for chronic conditions (99). A study of comparative effectiveness of pharmacologic interventions for knee OA utilizing a systematic review and network meta-analysis revealed: for pain, all interventions outperformed oral placebo; for function, all interventions except intraarticular corticosteroids were superior to placebo; and for stiffness, treatments did not differ from each other (100). Apparent superiority of some intraarticular treatments may relate to integrated intraarticular placebo effect. Limitations of published studies included lack of long-term data, inadequate reporting of safety data, possible publication bias, and few head-to-head comparisons.

Patients taking duloxetine for knee OA or chronic low back pain who had <10% reduction in pain after 4 weeks had limited possibility for eventually achieving moderate pain reduction by the end of 12 weeks (101).

Structural Outcomes of Pharmacologic Treatments

Structural outcomes were reported for strontium ranelate, recombinant human fibroblast growth factor 18, and glucosamine and chondroitin sulfate.

Strontium ranelate, 2 g/d, was associated with a beneficial effect on cartilage volume and bone marrow lesions in some regions, in a subset of the phase III Strontium Ranelate Efficacy in Knee Osteoarthritis Trial (102). Recombinant human fibroblast growth factor 18 was not associated with an effect on medial tibiofemoral cartilage thickness (the primary outcome), but secondary end points showed dose-dependent effects in a multicenter study conducted in Europe, North America, and South Africa (103). Glucosamine and chondroitin sulfate use was associated with less cartilage volume loss in an observational study (the Osteoarthritis Initiative), albeit not in multivariable analyses (104). In another study using Osteoarthritis Initiative data, glucosamine and chondroitin sulfate use was not associated with symptom relief or modification of disease progression (105). There was no evidence of structure benefit by MRI at 6 months or urinary CTX-II excretion from glucosamine in persons with mild-to-moderate knee pain in a randomized, double-blind, placebo-controlled trial (106).

Symptom Outcomes of Non-Pharmacologic Treatments

Symptom outcomes of non-pharmacologic interventions were reported for: neuromuscular exercise, quadriceps strengthening, weight reduction and maintenance, TENS, therapeutic ultrasound, stepped care strategies, cognitive behavior therapy for sleep disturbance, acupuncture, gait modification, booster physical therapy, a web-based therapeutic exercise resource center for knee OA; physical therapy for hip OA; and joint protection and hand exercises for hand OA. Reviews included meta-analyses of exercise, valgus bracing, ginger, electrical stimulation, and continuous and pulsed ultrasound.

Exercise and Physical Therapy

An intensive diet and exercise program (vs. diabetes support/education) prevented incident knee pain at year 1, in overweight adults with diabetes in the Action for Health in Diabetes study (107). Neuromuscular and quadriceps strengthening similarly improved pain and function but did not change the external knee adduction moment in the setting of moderate-severe medial knee OA with varus alignment in Australia (108). Neuromuscular exercise was more effective for non-obese persons with a varus thrust; quadriceps strengthening was more effective for obese persons without a thrust (109). Pressure-pain sensitivity, temporal summation, and pain were reduced with exercise in patients with knee OA (110). Booster sessions with a physical therapist did not influence pain, function, or home exercise adherence in patients with knee OA (111). A systematic review and meta-regression analysis of RCTs suggested that single-type exercise programs were more efficacious than programs including different exercise types for knee OA (112). In participants with knee OA in a 12-week exercise program, among tibiofemoral and patellofemoral MRI abnormalities, severity in patellofemoral cartilage integrity was associated with less improvement in physical function and strength and severity of patellofemoral osteophyte formation with less improvement in strength (113).

A program of exercise therapy and education (vs. only education) was associated with later total hip replacement (114). Hip physical therapy did not result in pain or function improvement vs. sham treatment for painful hip OA (115). For hand OA, joint protection (vs. no protection) was effective by 6 months using OARSI/OMERACT responder criteria; hand exercise (vs. no exercise) was not effective (116). Home-based exercise for hand OA was well tolerated and improved function, grip strength, pain, and fatigue in an RCT (117). In another RCT, group exercise followed by home exercise for hand OA was well tolerated but was associated with small short-term improvement only on self-reported measures (118). Short-term night time distal interphalangeal joint splinting was associated with reduced pain and improved extension in a radiologist blinded, non-randomized, internally controlled study (119).

Behavioral Therapy

Weight reduction in obese patients with a 1-year maintenance improved knee OA symptoms, irrespective of type of maintenance program (120). In a double-blind RCT with active placebo, cognitive behavior therapy for insomnia in knee OA reduced sleep maintenance insomnia and pain (121). In a secondary analysis of data from the Lifestyles trial, short term sleep improvement predicted long term improvements for sleep, pain, and fatigue that were not attributable to psychological benefits in persons with insomnia and joint pain (likely due to OA) (122). A web-based therapeutic exercise resource center was feasible and efficacious in improving pain, function, and self-efficacy in knee OA (123). Telephone reinforcement and negotiated maintenance contracts positively affected physical activity behavior in OA (124).

Complementary and Alternative Therapy

Studies of complementary and alternative therapies included a randomized clinical trial, sham-controlled, revealing no additional benefit of TENS over education and exercise for knee OA (125). In another RCT, double-blind and sham-controlled, therapeutic ultrasound provided no additional benefit over exercise training in persons with knee OA (126). Neither laser nor needle acupuncture conferred benefit over sham intervention for pain or function in patients older than 50 years with moderate or severe chronic knee pain (127). A systematic review and network meta-analysis suggested that pulsed ultrasound is more effective than control intervention (i.e., sham or blank) for pain relief and function improvement in knee OA (128). In a meta-analysis of randomized placebo-controlled trials for OA in any joint, ginger was modestly efficacious and reasonably safe for symptomatic OA, with evidence of moderate quality based on the small number of participants and frequently inadequate intention-to-treat analyses (129).

Biomechanical Intervention

Among biomechanical interventions, a meta-analysis of RCTs suggested valgus bracing was associated with small to moderate improvement in knee pain (130). A Cochrane review concluded evidence for an effect of bracing for medial knee OA on symptoms, function, and quality of life was inconclusive; evidence was moderate for a lack of effect from lateral wedge vs. neutral insoles on symptoms and function (131).

Structural Outcomes of Non-Pharmacologic Treatments

Only one study emerged in this category. In a randomized clinical trial in persons with painful patellofemoral OA, a patellofemoral brace reduced patellofemoral but not tibiofemoral bone marrow lesion volume and pain at 6 weeks (132).

Prevalence and Impact of OA

In 2 large referral centers in Cameroon, individuals tended to present for care at later stages of disease; among 148 with knee OA, moderate to severe disease was present in 36%, bilateral and bicompartmental disease in 38.5%, and bilateral tricompartmental disease in 14.2% (133). Pain and radiographic severity did not correlate (133). Another report described the prevalence of lumbar OA in Chinese adults (approximately 9%), with no difference between urban, suburban, and rural populations; higher prevalence was linked to gender (female), age, obesity, physical work, work posture, exposure to vibration during work, and duration of sleep (134). In the Framingham Osteoarthritis Study, the age-standardized prevalence of radiographic hip OA was 19.6%; prevalence of radiographic (but not symptomatic) hip OA was higher in men (135). The prevalence of OA features by ultrasound was described in participants from the Newcastle thousand families birth cohort (136). In a large sample of residents of Malmö, Sweden, 25% had radiographic knee OA and 15% had either symptomatic or clinically defined knee OA (137).

In the Global Burden of Disease 2010 study, of 291 conditions, hip and knee OA was ranked as the 11^th highest contributor to global disability and as the 38^th highest in disability-adjusted life years. While the global age-standardized prevalence of knee and hip OA did not change from 1990 to 2010, years lived with disability increased from 10.5 million in 1990 to 17.1 million in 2010 (138).

In cardiovascular populations (U.K. family practices), the comorbid addition of OA was associated with incrementally poorer physical health (by SF-12 Physical Component Summary) but interactions were less than additive (139). Patients with symptomatic OA at multiple sites (from the Genetics Arthrosis and Progression study and the Osteoarthritis Care Clinic study) who consulted health care providers for their OA were not at higher risk of death than the general population (140).

Health Services Research

A study of national cohorts of individuals with knee OA assembled using data from the 2003, 2006, and 2009 waves of the Medicare Current Beneficiary Survey revealed an increase in opioid prescribing between 2003 and 2009 (141). Correlates of opioid use included gender (female), functional limitation, poor self-reported health status, chronic obstructive pulmonary disease, and musculoskeletal disease apart from OA (141). In patients attending outpatient medicine and specialty clinics in the U.S., 16% rejected DMOADS under all conditions, 5% did not want to perform subcutaneous injections and would only consider DMOADs under the best-case scenario, 20% were willing to take DMOADS under the best-case but would start rejecting them as risk increased, and 59% would accept DMOADs across all scenarios (142).

Duloxetine (vs. comparators celecoxib, diclofenac, naproxen, hydromorphone, and oxycodone) was cost effective for 55-year old pts with OA and more so in older patients and those at greater risk for adverse effects in Canada (143). Celecoxib with a proton pump inhibitor was cost effective for OA (vs. diclofenac with a proton pump inhibitor) in an analysis within the Swedish health system using an adaptation of the NICE OA model (144).

A systematic review revealed well-developed, feasible indicators of quality of care for OA which could be implemented in primary care (145). Sponsored by the European League Against Rheumatism and the European Community, patient-centered and consensus based standards of care for OA were developed using systematic review and a 3-round Delphi process towards a goal of harmonizing treatment and care of OA within Europe (146).

Another report demonstrated a substantial increase in OA-related costs resulting from expanding total knee replacement eligibility criteria in the U.S., suggesting the need for more effective nonoperative therapies (147).

Special Reports – Research Priorities

A EULAR ad hoc expert committee concluded that research priorities include: predictors of OA progression (especially to enable stratified intervention); better understanding of mechanisms of OA pain; improved understanding of tissue communication; developing concepts of, and consequently interventions for, early OA; and a need for new therapeutic strategies (pathology-targeted and optimal combinations) (148) The recommendations of this report are summarized in Table 2.

TABLE 2.

OA Research Priorities, Adapted from a Report from a EULAR ad hoc Expert Committee (148)

Epidemiology	Progressive OA disease course (mechanisms and risk factors, prediction tool for clinical and research purposes, progressor phenotype)
	Early phases in OA development (criteria to define early or pre-OA)
	Criteria to diagnose and classify generalized or multisite OA
	Multidimensionality of OA outcomes (pain, function, participation, performance)
	Foot OA
Pathogenesis	Tissue communication in OA
	Non-cartilage articular pathology
	Mechanisms by which comorbidities influence OA process
	Joint trauma and subsequent repair
	Pathology at earliest stages of OA
	Relationship between pain and structure
Imaging and biomarkers	Performance metrics of imaging and other biomarkers
	Relevance of biomarkers to a broad range of domains
	Defining predictors of progression, especially those that aid targeted interventions (including combining imaging, biomechanical, biochemical biomarkers)
Therapy	Mechanisms of OA pain (origin of pain, its interrelation with other aspects of disease, in order to identify novel targets for pain management)
	Individualized or pathology-targeted therapies (advance knowledge of phenotypes of OA to target specific pharmacological and non-pharmacological therapies)
	Optimal combination therapy strategies (using complex intervention designs comparing monotherapy as well as specific combinations of pharmacological, non-pharmacological, and mixed approaches)

A prioritized research agenda about OA management was developed for the Patient-Centered Outcomes Research Institute; this report describes prioritized evidence gaps and top-ranked research priorities as well as recommended study designs or suggested analyses to address these gaps and priorities (149).

Conclusion

While much was accomplished in the past year, this review also illustrates directions for further work. Besides the obvious need to develop novel pharmacologic and non-pharmacologic interventions, future effort might include studies concerning: the relationship between physical activity and OA disease progression, modifiable determinants of physical activity behavior; interventions targeting physical activity, knee confidence, knee buckling; role of falls and fracture in OA, risk factors for falls and fracture; definitions of early OA, prevention/intervention efforts at pre- or early disease stages; identification/delineation of clinical phenotypes; the role of muscle function and quality in OA; potential disease-modifying effects of non-pharmacologic interventions; risk factors for incident and progressive hand OA, foot OA, hip OA; long-term pain outcome; and true disability outcome.