New Findings in Osteoarthritis:
 Etiology and Treatment

Publication Date: December 1, 2006 - Expiration Date: December 1, 2008

Supported by an educational grant from

Author:

Thomas A. Gossel, RPh, PhD
Professor of Pharmacology and Toxicology, Emeritus
Ohio Northern University
College of Pharmacy
Ada, Ohio

Statements made in this program have not been evaluated by the Food and Drug Administration. Discussion of published or investigational uses of drugs outside of approved labeling is offered for educational purposes only, and the sponsor and publisher of this program do not endorse such off-label use. Nutritional products discussed are not intended for the prevention, diagnosis, treatment, or cure of any disease.

INTRODUCTION

Arthritis is the most common chronic pathology of older adults in the United States, ranking at or near top prevalence for middle-age adults as well.^1,2 A survey of women under 45 years of age showed that 2% of the group had radiographic evidence of osteoarthritis (OA). The incidence increased to 30% between ages 45 and 64 years, and to 68% for women older than 65 years.³ Because of the increasing longevity of the United States population, the socioeconomic impact of OA will become even more significant in the near future.^4,5 By the year 2020 the worldwide population of persons over age 50 will be double that of 1990, with increased spending on diagnosis, therapy, adverse effects prevention, and loss of productivity.⁶ Today, OA costs the nation’s economy more than $86 billion annually: $51.1 billion in direct medical-related costs, and $35.1 billion in indirect costs or lost wages.^7,8

This monograph discusses current medical thinking regarding OA with emphasis on the inflammatory component. It discusses the role of nonsteroidal anti-inflammatory drugs (NSAIDs) in OA, including those available without a prescription. It also provides specific information to use when counseling patients or their caregivers on OA and its therapy.

OSTEOARTHRITIS

OA is a chronic and progressive articular pathology characterized by deterioration of cartilage along with formation of new bone stimulated by the disease. The condition is characterized by the gradual development of joint pain, stiffness, and limitation of motion.⁹ As degeneration of cartilage is the most prominent pathologic change, the terms degenerative arthritis and degenerative joint disease are often used to describe OA and are used interchangeably with osteoarthritis. The presence of synovial inflammation has been convincingly demonstrated in OA, at least during some phases of the disease, and a number of observations suggest a role for joint inflammation in the pathogenesis of pain.¹⁰ Even though inflammatory changes may be absent early on,¹¹ inflammation will cause many, if not most, of the clinical problems later on.^12,13

OA is the most common of all the arthritides.¹⁴ Because arthritis is ubiquitous in the elderly and is rarely associated with mortality, it is often viewed as a benign condition of aging. However, arthritis is hardly benign.

The most prevalent form of OA is primary or idiopathic (ie, no identifiable cause).⁶ In primary OA, no specific traumatic or other predisposing factor can be identified. This contrasts with the less common form, secondary OA, in which there is a known cause (eg, trauma, anatomic abnormality, metabolic disease) that accounts for the changes.^3,15

Osteoarthritis pain accounts for a disproportionate share of lost productive work time.¹⁶ Even mild or moderate pain, especially if unrelenting, may negatively impact one’s physical and emotional state.^17,18 OA of the knee or hip in particular can severely compromise activities such as walking, climbing stairs, and using the bathroom.¹⁹ This toll on quality of life extends across every age and stage of life, and has been demonstrated for nearly every form of pain, including persistent pain of OA.¹⁷ The World Health Organization reports that persons with persistent pain are four times more likely than those without pain to suffer from depression or anxiety, and twice as likely to experience difficulty in working.²⁰

Risk Factors
Risk factors for OA are well documented and include both systemic and local factors. The predominant factors are advancing age and female gender, although a genetic component is apparent. Nutritional deficiencies may be present in some groups.²¹

Increasing age is the strongest risk factor for OA. This is most likely related to age-associated biologic changes, such as reduced responsiveness of chondrocytes to growth factors that stimulate joint repair, an increase in ligamentous laxity that may render older joints unstable and thus more susceptible to injury,¹⁵ and failed shock absorption or protection of the joint with age.

A role for genetic factors in development of OA in certain joints more than others (ie, hands and knees in women, but not the hip) has been suggested for more than a half century. Based on clinical findings, Heberden’s nodes (small, hard nodules on the distal interphalangeal joints of the fingers) were demonstrated to be 10 times more likely to occur in women than in men, and mothers and sisters of women with Heberden’s nodes were two to three times more likely to be affected than the general population.^22,23 Spector and coworkers have proven an increased correlation of radiographic OA of hands and knees among identical twins when compared with fraternal twins.²⁴ They surmised that the genetic influence on hand and knee OA ranged from 39% to 65%.

Obesity is a well-established risk factor for OA of the knee but not the hip. Framingham Study cohort data suggest that obese individuals, particularly women, are more likely to develop knee OA; similarly, weight loss appears to reduce the incidence of symptomatic knee OA.²⁵

Known risk factors for developing OA are summarized in Table 1.

Table 1. Risk factors for osteoarthritis^3,43,55,135 

   -Increasing age
   -Congenital/developmental defects
   -Obesity
   -Female gender
   -Repetitive (eg, occupational) stress
   -Major joint trauma
   -Sports injuries
   -Crystalline deposit disease
       Calcium pyrophosphate
       Hydroxyapatite
   -High bone mineral density
   -Previous inflammatory joint disease
   -Peripheral neuropathy
   -Osteoporosis (controversial)
   -Metabolic/endocrine abnormalities
       Acromegaly
       Wilson’s disease

Pathophysiology
Multiple factors contribute to development of OA, including the combined action of biomechanical, biochemical, and immunologic forces, along with inflammation.²⁶ These repetitive stresses induce injury to the articular cartilage.²⁷ Of the involved tissue systems, the most relevant to an understanding of the pathology of degenerative arthritis are cartilage, synovial fluid, and subchondral bone.²⁸ The earliest radiographic signs of OA include irregularities in the surface of articular cartilage, which cause joint space narrowing. Initial irregularities result from release of intracellular proteolytic enzymes, including collagenases, which initiate the enzyme cascade; matrix metalloproteinases (MMPs), which degrade both proteoglycans and collagen; and aggrecanases, which along with MMPs, degrade aggrecan. Chondrocytes proliferate in response to protein degradation,²⁹ and produce macromolecules including collagen, hyaluronan, and other proteins. OA results when synthesis (ie, anabolism) and degradation (ie, catabolism) processes involving these structural components become unbalanced, and the rate of catabolism exceeds cartilage synthesis.¹ Progression of the degenerative process is thus dynamic, involving continuous cycles of degradation and repair, which may help explain its chronic course.

Because of its elastoviscosity, synovial fluid stores mechanical energy within its molecular network and releases it in the form of heat through movement of the network.³⁰ The rich hyaluronan content of synovial fluid permits it to function as both lubricant during slow movement and elastic shock absorber during rapid movement.³¹ Synovial fluid also serves as a medium for delivering nutrition and conveying cellular signals to articular cartilage. Its quantity increases in response to inflammation, resulting in a proportional decrease in hyaluronan content with consequent impediment of its functions.²⁸

Subchondral bone undergoes substantial remodeling in OA, including increased density (sclerosis), formation of cyst-like bone cavities, and proliferation of regenerative cartilage both within the bone and on its surface.³² Sclerosis of subchondral bone leads to formation of osteophytes, typically on the periphery of the joint, appearing as an extension of the joint surface. The presence of these bony outgrowths restricts motion and causes pain. It is not known whether osteophytes arise before, during, or after cartilage degeneration, or whether one contributes to the degradation of the other.

As noted earlier, the frequency of OA increases with age. This may be due to increased susceptibility of cartilage to damage resulting from abnormal biomechanics, which then progresses to excessive stress or defects in the matrix components and alterations in the normal structure. Repetitive stress and impact loading are more damaging than frictional wear. Early in the disease, synovitis is usually minimal, but may contribute to joint damage as inflammation progresses. The result of microfractures and change of hardness of bone and ability of cartilage to absorb stress have also been suggested to be the primary mechanism in OA.

Prostaglandins and pain.—Of significance to the pathogenesis of pain associated with OA is an understanding of the biologic actions of prostaglandins, including their mechanisms of synthesis and inhibition. Prostaglandins are fatty acid derivatives that contribute to a variety of physiologic and pathologic processes, including inflammation and pain. They are synthesized from arachidonic acid, a 20-carbon chain polyunsaturated fatty acid supplied from dietary sources or derived by conversion from linoleic acid. Arachidonic acid is a component of cellular membrane phospholipids and released via hydrolysis by cellular phospholipases as a result of mechanical and physical stimuli or other chemical mediators.³³ Arachidonic acid metabolites (eicosanoids) are synthesized by catalytic activity of cyclooxygenase (prostaglandins and thromboxanes) and lipoxygenase (leukotrienes). Eicosanoids can arbitrate virtually every step of the inflammation cascade.

Prostaglandins are potent mediators of pain. They do not cause pain on their own, but rather sensitize nerve endings (pain receptors) to mechanical and chemical stimulation.^34,35 Prostaglandin E₂ (PGE₂) and prostaglandin I₂ (PGI₂) are particularly potent pain inducers. Other prostaglandins may contribute to pain by their interaction with additional chemical mediators. Inhibition of prostaglandins with pharmacologic agents, as described subsequently, reduces the severity of inflammation and consequently, OA pain.³⁴

Cyclooxygenase.—Cyclooxygenase (prostaglandin endoperoxide synthase; COX) modulates a number of physiological functions including the process of inflammation. Two isoforms, called COX-1 and COX-2, have been identified to date,^35,36 and serve different functions. COX-1 is present in most normal cells throughout the body and is synthesized continuously to stimulate synthesis of prostaglandins that regulate basal cellular activity. Notable among these reactions are platelet aggregation and maintenance of homeostasis of the vasculature, and of the gastrointestinal mucosa and renal system. In contrast, COX-2 is found in fewer tissues. It appears to be constitutive in some tissues, including the brain, reproductive organs, kidney, and placenta during late gestation.³⁷ COX-2 is expressed (induced) rapidly and more selectively than COX-1 in response to local inflammatory stimuli and released in tissues, where it initiates the characteristic response of inflammation and pain.^11,34,38 The two isoforms differ by a single amino acid at position 509 of cyclooxygenase (COX-1, valine; COX-2, isoleucine).

Clinical Features and Diagnosis
The typical patient with OA is middle-aged or elderly, and complains of pain in one or more of the sites shown in Table 2. Symptoms may be absent in early disease since articular cartilage is devoid of neural intervention.³⁹ Symptoms (Table 3) are often insidious in onset.

^*Number of respondents varied from 336 to 339 (total sample n = 339) for the items above; responses are not mutually exclusive.

Table 3. Clinical features of osteoarthritis¹ 

Symptoms
 -Joint pain
 -Morning stiffness lasting ≤20 minutes
 -Joint instability or buckling
 -Loss of function

Signs
 -Bony enlargement of affected joints
 -Limitation of range of motion
 -Crepitus on motion
 -Pain with motion
 -Malalignment and/or joint deformity

Pattern or joint involvement^*
 -Peripheral: DIP, PIP, MCP joints; knees; hips
 -Axial: cervical and lumbar spine
_____________
^*Disease with multiple joint involvement is a subtype of osteoarthritis; most commonly, osteoarthritis affects the  hands, hip, knees and/or spine.

DIP = distal interphalangeal; PIP = proximal interphalangeal; MCP = metacarpophalangeal

Since there are no specific diagnostic tests for OA, the clinical diagnosis will be based on patient history, physical examination findings, and radiographic features.⁴⁰ OA is not a systemic disease so it is not associated with generalized and prolonged morning stiffness, fever, or loss of appetite. Patients may experience stiffness in involved joints upon awakening, but compared with rheumatoid arthritis (RA), stiffness does not persist as long, usually only about 20 minutes. Crepitus (“crackling” sound upon moving the joint) is common in both OA and RA due to the irregularity of opposing cartilage surfaces. Other similarities and differences that differentiate OA from RA are shown in Table 4. Attempts to discriminate between OA and RA with verbal description often prove unsuccessful.⁴¹ Signs of OA may also include bony enlargement of affected joints, limitation of range of motion, and malalignment and/or joint deformity.⁴² Pain is characterized as a deep, aching or throbbing sensation that intensifies with motion and improves with rest.⁴³ The usual response may be interspersed by activity-related episodes of sharp, stabbing pain.⁴¹ It is usually intermittent and often mild, but can become persistent and severe. Severe pathology may involve pain at rest or during nighttime hours. Psychological variables are the strongest predictors of pain severity and include anxiety, depression, hypochondriasis and negative emotion.^44,45

Physical examination should include careful assessment of the affected joints. Alternative conditions should be ruled out. Intensive joint inflammation may signal that the arthritis is advanced and erosive. Special attention should be given to the patient’s history of present illness to differentiate it from RA. A joint that appears hot, reddened, and markedly swollen suggests crystal arthropathy, such as gout or infection.⁴²

In primary OA, no specific clinical abnormalities are noted. Laboratory tests including erythrocyte sedimentation rate (ESR), blood chemistry and morphology studies, and urinalysis are usually within normal limits. In contrast to RA, OA is associated with a negative rheumatoid factor. Analysis of synovial fluid from involved joints usually reveals a high viscosity and mild leukocytosis (<2000 WBC/mm³) with predominantly mononuclear cells.⁴²

Radiologic evaluation can confirm the diagnosis of OA although the outcome can be nonspecific because radiologic changes may not be apparent in early or mild forms of the disease. Disease progression will show a loss of joint space and presence of osteophytes or new bone formation. With progression, joint deformity and subluxation (incomplete or partial dislocation) can occur.⁴² 

MANAGEMENT

Public health data that characterize arthritis as the leading cause of functional limitation, physician visits, and missed workdays disavow any suggestion that arthritis is a benign condition. Pain is the reason most patients with OA seek professional care.^39,46 Medical and rehabilitative care ideally includes effective means to treat the condition and reduce the pain, and prevent or correct physical impairments and associated functional limitations of the disease. Primary care physicians, rheumatologists, physiatrists (physicians who specialize in physical medicine), physical and occupational therapists, surgeons, orthopedists, nurses, and pharmacists, therefore, can all contribute to the care of arthritis patients who have varying levels of need and functional limitation.      

Ongoing medical research has greatly increased understanding of the pathogenesis of OA and treatments continue to be created and refined.²⁸ Despite promising research, at present, there are no drugs capable of modifying the structural damage of OA or reversing its pathogenesis.⁴⁷ In spite of current limitations, patients should be encouraged that, while their disease is not curable, pain and associated symptoms can be effectively controlled in almost all cases.⁴⁸ Effective drug and lifestyle changes can reduce pain and improve physical functioning.⁴⁹ Controlling pain will help achieve the other goals of therapy.

Individual response to pain management therapy varies greatly. Some patients will improve greatly following standard pharmacologic and nonpharmacologic intervention, while others with inflammation and joint injury of equal magnitude continue to experience pain after receiving identical treatment. Successful management depends on an accurate diagnosis including extent of joint involvement, along with therapy that is individualized to the patient and most likely includes physical and occupational therapy. Patient education is vital to a successful outcome. Measures that maintain adherence to a regime, such as keeping a personal diary or receiving social support from friends or family members, are thought to improve long-term outcome.⁵⁰ Surgery may be an option for extensive involvement. 

The goals of therapy for OA are to^17,18:

• manage the severity and duration of pain and other symptoms related to the disease;
• educate patients, along with caregivers and other family members to encourage patients to be proactive in managing the disease;
• retard disease progression to preserve joint motion; and,
• promote patient independence and improve their quality of life in all activities of daily living.

Numerous factors can modify the patient’s progress or alter the expected time frames; therefore, the goals should be revised as often as needed to maximize patient benefit and comfort. The presence or absence of comorbid pathologies, concomitant medications, and allergies may also affect the general approach to treatment.

Patient Education
Patients with OA, especially early in the disease process, may attribute symptoms to normal “wear-and-tear” conditions associated with sports or their occupation, or with advancing age; in either case, they do not seek medical care. Or, they may deliberately avoid activity that exercises affected joints excessively. A vicious cycle can develop to inactivity whereby failure to exercise a joint over prolonged periods can exacerbate the disability resulting from progression of arthritis. The tendency of many people to become less physically active with increasing age adds to the problem in that people’s thinking links increased “wear and tear” with exercise and movement, or conversely, lessened degeneration with rest and inactivity. Only 24% of people with arthritis receive sufficient physical activity to achieve health (Figure 1). Indeed, arthritis is the major reason why elderly persons are not active or limit their activity.⁵¹ The outcome of numerous studies in persons with OA has demonstrated the benefit of movement and activity on all joint tissues – cartilage, bone, muscle, and ligaments.⁵²

Figure 1. Extent of physical activity reported by arthritic patients. Only 24% of patients report and achieve levels of physical activity that are recommended for health.⁵¹

Education is therefore the first step in conservative treatment of OA.^53,54 Patients should thoroughly understand the extent and degree of involvement, reasons for the approach to management, and expected prognosis of their disease.⁵⁵ Reasons for and importance of nonpharmacologic therapy should be stressed continually.⁵³ Numerous self-help primers have been prepared to aid the education process,⁴² and are available through local book stores and libraries. The Arthritis Foundation can be contacted (www.arthritis.org; 1-800-283-7800) for guidance. Local agencies that respond to patients who suffer from OA and/or other rheumatoid afflictions may be identified in telephone directories. Arthritis self-help courses taught by local allied health professionals teach patients how to manage their disease. Participation in such courses can reduce pain and improve quality of life.⁵⁶

Physical and Occupational Therapy
Physical and occupational therapy is designed to help patients regain and maintain range of motion, reduce spasms and relieve pain, and is a critical component of OA disease management.⁵⁷ Effective therapy improves muscle strength and teaches more efficient joint management and energy conservation. Such interventions include joint protection, training in activities of daily living, and functional exercise.³⁹ The goal is to maintain and improve function of affected joints,⁵⁸ and minimize the risk of other chronic conditions among individuals with OA.^59-63

The National Arthritis Action Plan⁶⁴ and Healthy People 2010⁶⁵ underscore the importance of exercise in persons with arthritis. Exercise is most effective if it includes strengthening muscles involved in activities a person normally performs each day. Range-of-motion exercise alone does not strengthen muscles. Isotonic exercise strengthens muscles, but not through a range of motion. The patient needs to perform both isotonic and range-of-motion exercises to improve joint function. Numerous clinical trials have consistently demonstrated the efficacy of isokinetic or isotonic strengthening (ie, strengthening that occurs when a person flexes or extends the knee against resistance) in reducing pain and improving function. Low-impact aerobic exercise is also effective in lessening pain. The involvement of a physical or occupational therapist will maximize patient response and greatly improve quality of life.^43,66

Thermotherapy.—There are no large, controlled clinical trials that demonstrate therapeutic benefit of applying heat to arthritic areas. Nevertheless, practical experience suggests that heat can lessen stiffness and provide symptomatic pain relief of OA. Heat can be applied for 15-20 minute periods using electric heating pads, hot towels, hot-water bottles, or heat packs.⁶⁷

Paraffin wax baths can also be used to apply heat uniformly to the hands, elbows, or feet.⁶⁷ By dipping the affected joint into the warm paraffin, a soft coating of warm wax will release its warmth slowly and uniformly. Warm paraffin baths should not be used over open wounds because hardened wax may contaminate the wounds and the wax can be difficult to remove. Patients with OA of the hands can be advised to purchase the thin, disposable plastic gloves commonly used by food workers or hobbyists and wear them before dipping their hands into the warm wax. These inexpensive, single-use gloves are available from commercial food suppliers and/or hobby shops.

Prolonged heat or temperatures above 110^o F should be avoided, regardless of the method of thermotherapy, to prevent skin damage. Commercially manufactured wax-bath units are available with temperature-control capacity.

Impaired circulation, poor thermal regulation, edema in the area being treated, open wounds, or areas where topical rubefacients and counterirritants have been applied may preclude the use of heat.²⁵ Patients should be evaluated for impaired circulation. If the vasculature is impaired, it may not be able to dilate normally in response to heat, thus predisposing patients to burns. Signs of compromised circulation include poor nail quality, decreased ambient skin temperature, ulcerations, and thin skin.

Cold therapy is very useful in the initial treatment of certain acute painful musculoskeletal syndromes such as spasm and soft tissue injury.⁵⁴ Its value in OA has been debated.²⁵

Support Braces
OA pain in joints at the base of the thumb may be relieved by wearing an elastic wrist or hand support brace. These units immobilize the joints while permitting the patient to proceed with activities of daily living. Braces also reduce inflammation, protect weak joints, preserve anatomic alignment, and enhance functions.²⁵ Carpal tunnel wrist supports, worn nightly, are useful for patients with inflamed arthritic wrists who develop a compression median neuropathy because of increased intracarpal canal pressures.

Patients with extensive OA of the hand may also benefit by wearing two-way stretch gloves, either nylon or spandex, overnight and other times when working with the hands. Worn in early disease stages, these gloves can actually prevent arthritic nodal development or reduce nodal size, and enhance function in later stages.^68,69

Ready-to-wear supports and braces contain detailed sizing charts within the package. Healthcare professionals should familiarize themselves with the methodology for sizing so they can assist patients in their proper selection. Persons who suffer from severe occlusive arterial pathology, such as Raynaud’s disease, and others with dermatologic infections or allergies to latex rubber in the products should not use wrist supports without their physician’s approval. The supports may decrease arterial blood supply into the tissues of the hand and result in increased symptom severity or onset of other problems. Anyone with these concerns should be advised to consult his or her physician about the products.

Nonpharmacologic means alone may not relieve OA symptoms. Pharmacologic therapy, unless specifically contraindicated, may be incorporated into management of the disease. Generally, patients who do not meet the criteria for exclusion from self-care (Table 5) can self-treat their pain. When used as directed, nonprescription analgesics are safe and effective for self-treatment of mild to moderate pain of OA.⁷⁰

Table 5. Exclusions to self-treatment of osteoarthritis and other musculoskeletal pain^*76 

·         Pain persisting longer than 2 weeks

·         Pain with nausea or severe vomiting

·         Weakness in any limb

·         Pain in a red, hot, or swollen joint (with or without fever)

·         Suspected fracture

·         Pelvic or abdominal pain (other than dysmenorrhea)

·         Suspected pregnancy (last menstrual period >6 weeks ago)

·         Increased intensity of pain or any change in the character of pain

__________________

^*Management of chronic musculoskeletal pain is appropriate with OTC medications following evaluation and under the supervision of a physician.

Nonsteroidal Anti-inflammatory Drugs
NSAIDs are the most commonly used pharmacologic agents worldwide and the most frequently prescribed analgesics for management of OA pain of mild to moderate intensity.^71-73 It is estimated that more than 30 million people worldwide take NSAIDs daily for treatment of pain,⁷⁴ with annual worldwide sales exceeding $20 billion.⁷⁵ An estimated 30 billion doses of nonprescription NSAIDs alone are consumed each year in the United States⁷⁶ by 13 million individuals who suffer from arthritis.³³ All NSAIDs possess analgesic, antipyretic, and anti-inflammatory action.⁷⁷ Smaller doses produce analgesia; higher doses are anti-inflammatory.³⁹ Since pain relief is proportional to dose and the drugs reach a ceiling for maximum activity, analgesic tolerance to NSAIDs does not appear to occur. Moreover, NSAIDs do not alter the perception of sensory modalities other than pain.⁷⁷

MANAGEMENT OF OSTEOARTHRITIC PAIN WITH NSAIDs  

Their major mechanism of action is through inhibition of COX (Figure 2), the enzyme responsible for synthesis of prostaglandins from precursor arachidonic acid.^77,78 The non-aspirin, traditional NSAIDs (eg, ibuprofen, naproxen) differ from aspirin. Aspirin binds covalently with both COX-1 and COX-2 to result in irreversible inhibition of enzymatic activity. The vast majority of traditional NSAIDs are organic acids and therefore are reversible, competitive inhibitors of COX. As noted earlier, COX is dispersed throughout the body and involved in a wide variety of physiologic processes.

Figure 2. Mechanisms of action of NSAIDs

With discovery and characterization of COX-2 and its action to induce inflammation and pain, intensified research led to development of highly selective COX-2 inhibitors (eg, celecoxib, rofecoxib, valdecoxib; “coxibs”) (Figure 3). Coxibs have been demonstrated repeatedly to be at least equal to traditional NSAIDs in effectiveness. An advantage of selective COX-2 restriction is that the drugs provide anti-inflammatory and analgesic effects while, in theory, reducing gastrointestinal toxicity and platelet inhibition associated with COX-1 inhibition.^33,79 However, once considered a safer alternative to nonselective COX inhibitors, coxibs have come under intense medical scrutiny since emerging evidence has implicated them in adverse cardiovascular events.^80-82 Rofecoxib and valdecoxib have subsequently been withdrawn from therapy in the United States. All traditional NSAIDs are nonselective in that at therapeutic concentrations, they inhibit both COX-1 and COX-2 isoforms to similar extent, although individual NSAIDs inhibit COX-2 to various extent from one another. 

Figure 3. Inhibition of COX-1/COX-2 by NSAIDs

The 1999-2000 National Health and Nutrition Examination Survey (NHANES) revealed that 20% of US adults used nonprescription or prescription non-narcotic analgesics nearly every day for at least a month at some point during their lifetime.⁸³ Moreover, 14% of adults were currently using analgesics (8% aspirin, 3% non-aspirin NSAIDs; 3% acetaminophen) frequently. Three-fourths of aspirin users (which includes its use in cardioprotection), 46% of non-aspirin NSAID users, and 63% of acetaminophen users had used them for one or more years. Seven percent of frequent analgesic users took two or more analgesics nearly every day. The survey concluded that, because frequent non-narcotic analgesic use, especially of drugs available without a prescription, is so prevalent in the United States, healthcare professionals should routinely monitor nonprescription, as well as prescription, analgesic use in their patients to prevent adverse drug effects and inappropriate use.

Another survey evaluated the frequency and indications for OTC NSAID use, and extent to which the public is aware of the products’ safety.⁸⁴ Investigators analyzed the responses of 9,062 patients who participated in earlier surveys. Seventeen percent of respondents in the first survey used NSAIDs, with 38% of users taking both prescription and OTC products. Forty-six percent of exclusive OTC product users believed OTC NSAIDs were safer, while 56% of exclusive users of prescription NSAIDs believed they were safer. Sixty percent and 29% of exclusive OTC users were neither aware of nor believed they were at risk for side effects from NSAIDs, respectively. Twenty-six percent of respondents used more than the recommended dose and 22% believed warning symptoms would always precede NSAID-induced complications. In a parallel survey, 83% had used an OTC NSAID within the past year; 15% reported daily use. Forty-nine percent were not concerned about potential adverse effects. Thirty percent of respondents indicated there was less risk of adverse effects with OTC analgesics and 44% consumed more than the recommended dosage on the label. The investigators reported that OTC NSAIDs are widely used, often taken inappropriately in potentially dangerous doses, and users were frequently unaware of their potential for adverse effects. They urged strong educational intervention directed toward both patients and physicians.

All FDA-approved nonprescription analgesics for internal use are NSAIDs. Nonprescription analgesics are listed in Table 6.

^*Restriction to limit dosage in persons ≥65 years has been removed from product labeling.  Patients over 65 years should check with a physician before taking any medication.

Nonprescription versus Prescription NSAIDs
The likelihood of an adverse drug reaction or drug interaction is much greater in patients with a minor ache or pain who take the maximum dose of an OTC NSAID for one week than in a person with OA who takes a prescription anti-inflammatory dose of the same drug over many months, regardless of whether the drug was purchased with or without a prescription. This can greatly confound patients’ and healthcare professionals’ understanding of drug safety issues. For example, much of the professional data published on NSAID safety reflect actions that can be expected from doses that are clearly in the realm of prescription drug use rather than at levels representative of self-treatment and therefore do not directly apply to OTC product uses. At the same time, it must be emphasized that some patients will self-administer nonprescription NSAIDs at dosages and for durations that exceed the guidelines listed in their labeling. Long-term safety studies that employ conditions in the prescription domain, therefore, cannot be ignored totally.

Aspirin
Aspirin is widely used for relief of low-to-moderate intensity pain and long-term use does not lead to tolerance or addiction. The drug alleviates pain by its peripheral action to inhibit COX-2. Direct effects on the CNS also may be involved.⁷⁷

Adverse effects include gastrointestinal distress and decreased platelet aggregation.⁷⁷ Concurrent ingestion of food, milk, or antacids may help minimize local gastrointestinal effects; however, enteric-coated aspirin products should not be taken with milk or antacids since the tablet coating may be compromised. The non-aspirin NSAIDs  at OTC dosing typically cause fewer gastrointestinal effects and bleeding than aspirin.

Acetaminophen
Acetaminophen (paracetamol; N-acetyl-p-aminophenol; APAP) is the active metabolite of phenacetin, which was withdrawn from use as an OTC analgesic in the United States in the late 1970s, when it was implicated in analgesic-abuse nephropathy.⁸⁵ Although acetaminophen was first used as an analgesic in 1893, it did not gain in popularity until 1949, after it was identified as the major active metabolite of phenacetin. Acetaminophen is a weak NSAID peripherally, with its primary action to inhibit COX-2 centrally.⁷⁷ Acetaminophen differs from aspirin and the traditional NSAIDs by its poor antiplatelet and anti-inflammatory actions peripherally. Including acetaminophen in the same pharmacologic category as NSAIDs may be confusing to healthcare professionals whose classic pharmacology instruction taught that acetaminophen is without anti-inflammatory action.

Acetaminophen is generally well tolerated and produces milder gastrointestinal and renal effects than traditional NSAIDs. The risk of upper GI bleeding with doses >3,000 mg/day is higher than with lower doses.¹⁷⁸ It may produce hepatic toxicity when taken in excess of 4,000 mg/day for prolonged periods.⁸⁶ Toxicity is intensified in persons with hepatic disease or by alcohol ingestion,⁷⁷ and has been reported with therapeutic doses (1,000 mg 4 times daily for 4 days).⁸⁷ Intentional or unintentional overdose of acetaminophen has become the most frequent cause of acute liver failure in the United States, United Kingdom, and northern Europe.^88,89 It is the leading cause for calls to Poison Control Centers (>100,000/year) in the United States, accounting for more than 56,000 ER visits, 2,600 hospitalizations, and an estimated 458 deaths due to acute liver failure each year.⁸⁸ Recently, acetaminophen was strongly implicated as having an adverse effect on endothelial function causing blood pressure elevation.⁹⁰ Patients should be aware that numerous nonprescription and prescription products contain acetaminophen.

The effectiveness of acetaminophen versus NSAIDs in modifying inflammatory pain remains the subject of an ongoing debate. A recent comprehensive Cochrane review summarized the worldwide published data on acetaminophen for OA pain.⁹¹ It noted that acetaminophen is less effective than traditional NSAIDs for pain reduction, global assessments of efficacy, and improvement in functional status.

The 1996 American College of Rheumatology (ACR) treatment guidelines for OA listed acetaminophen as first-line pharmacologic therapy. This recommendation was modified in the updated 2000 (ie, current) guidelines in that the traditional non-aspirin NSAIDs were listed as an alternative initial approach to therapy.⁸⁶ The current recommendation was based on emerging evidence suggesting that traditional NSAIDs were superior to acetaminophen in management of OA-related pain,⁹² in view of the inflammatory etiology of OA. The current Canadian Consensus Guidelines recommend NSAIDs as treatment of first choice for moderate to severe OA.⁹³ Some patients still may prefer acetaminophen over traditional NSAIDs because of advertising claims that it causes fewer adverse gastrointestinal effects, even when traditional NSAIDs have superior activity on the inflammatory component (ie, cause) of their OA pain.⁵¹

OTC NSAIDs
Ibuprofen.—Following a decade of use restricted to prescription only, ibuprofen was switched to OTC status in 1984.⁹⁴ Since then, numerous clinical trials have demonstrated that ibuprofen is suited for use as a nonprescription analgesic. It is generally well tolerated for treatment of acute pain.⁹⁵ To illustrate, the safety of single doses of nonprescription-strength ibuprofen was evaluated by examining documented adverse effects reported in 2,579 patients representing 15 double-blind, randomized, controlled trials.⁹⁶ Eight hundred seventy-eight subjects received ibuprofen 200 mg or 400 mg, 849 received acetaminophen 650 mg or 1,000 mg, and 852 received placebo. The overall frequency of adverse effects was comparable: ibuprofen 2.4%, acetaminophen 3.2%, and placebo 2.1%. Upper GI upset ranged from 0.8%-0.9% of subjects in all groups.

Naproxen sodium.—Naproxen was transferred from prescription to nonprescription status in 1994 after 12 years of use.^94,97 In comparative studies, naproxen was considered the best tolerated, followed by ibuprofen and fenoprofen.⁸⁵ Efficacy and tolerability of naproxen have been substantiated over many years of clinical use. Therefore, it can be considered a first-line treatment for OA, other rheumatic diseases, and various pain states.⁹⁸ OTC products contain the sodium salt of naproxen because it dissolves more rapidly in gastric juice and is absorbed more quickly than the free chemical, resulting in a quicker onset of analgesic effect.

Ibuprofen vs. naproxen sodium.—Ibuprofen and naproxen sodium were compared for analgesic efficacy and safety in patients with OA of the knee.⁹⁹ In two identical multicenter, randomized, double-blind, placebo controlled, parallel-design clinical trials, patients aged ≥25 years were randomized to daily doses of naproxen sodium 660 mg (440 mg in patients ≥65 years), ibuprofen 1,200 mg, or placebo, for 7 days. Drug efficacy on seven criteria (pain at rest, pain on passive motion, pain on weight bearing, stiffness after rest [morning], day pain, night pain, 50-foot walk time [seconds]) was evaluated. Both NSAIDs were clinically effective in pain-relieving efficacy compared with placebo. Both treatments reduced the mean symptom score by 30%-45%, compared with 20%-25% reduction with placebo. In patients aged ≥65 years, naproxen sodium was significantly superior to placebo in relieving all symptoms except pain on weight-bearing; ibuprofen significantly reduced day pain only. For daily pain diary evaluations, both NSAIDs effectively reduced all six symptoms. There was a trend toward greater efficacy for nighttime pain relief with naproxen sodium compared with ibuprofen. There were no significant differences in adverse events between groups.

Choosing an OTC analgesic.—There are several points to consider when choosing an OTC analgesic. The first relates to pharmacokinetic differences in their half-lives. Aspirin, with a plasma half-life of 15 minutes, and ibuprofen and acetaminophen, with half-lives of about 2 hours, must be administered every 4-6 hours. Naproxen, with a half-life of nearly 14 hours, may be administered every 8-12 hours, making it well suited for persons with long-lasting pain. Nonetheless, the choice of agent is often empirical and depends on the individual’s response to therapy. Cost is an important factor, especially for agents that, with physician guidance, will be taken over an extended period.

Numerous clinical investigations have attempted to identify advantages of one analgesic over the others. For example, a mail survey evaluated the treatment preferences of 1,799 patients with OA, rheumatoid arthritis, or fibromyalgia.¹⁰⁰ Approximately one-third of patients who had taken acetaminophen believed it to be moderately or very effective, while the others indicated that acetaminophen was not effective or was only slightly effective. When both efficacy and adverse effects were considered together, 25% of the patients had no preference, 14% preferred acetaminophen, and 60% preferred a traditional NSAID. Similar patient preferences were noted across all categories of patients and disease severity. The study authors concluded that “…there would hardly ever be a reason to recommend acetaminophen over NSAIDs, since patients generally preferred NSAIDs and fewer than 14% preferred acetaminophen.”

A meta-analysis of seven published articles that described treatment efficacy in 1,252 subjects with symptomatic hip or knee OA (752 on NSAID; 500 on acetaminophen) concluded that traditional NSAID therapy was statistically superior in reducing pain at rest and walking compared with acetaminophen.¹⁰¹ Drug safety, measured by discontinuation due to adverse events, was not statistically significantly different between NSAID and acetaminophen users.

NSAID Cardiovascular Safety
Following publication of the results of several large clinical trials that seemed to support an increased risk for cardiovascular toxicity for coxibs, the FDA concluded that there is a “class effect” for risk of increased cardiovascular events for all non-aspirin NSAIDs.¹⁰² The agency ruled that the professional labeling for all prescription NSAIDs, including coxibs and the traditional drugs, must be revised to include a boxed warning stressing the potential for increased risk of cardiovascular events and a Medication Guide be dispensed.^103,104 The boxed warning also includes the well-established risk of serious GI bleeding. Additionally, the labeling for all NSAIDs includes a contraindication for use in patients immediately following coronary artery bypass grafting surgery.

The FDA noted that ibuprofen and naproxen are currently available without a prescription for short-term use in acute pain and stated that these nonprescription products should continue to be available to consumers. It acknowledged that this is the first time that products with a boxed warning in their prescription labeling are available for nonprescription use. The FDA was convinced that scientifically sound data support the short-term use of usual doses of the nonprescription NSAIDs that are generally well below the maximum daily intake for the same ingredient and the duration of treatment, and are not associated with increased risk of serious cardiovascular events.¹⁰³ The benefit-versus-risk profile is favorable when OTC NSAIDs are used according to labeled instructions.

Naproxen has been identified as the single NSAID for which there may be an exception from the implication of increased risk for cardiovascular events.¹⁰³ Aisen and colleagues undertook a randomized, placebo-controlled study in patients with Alzheimer’s disease to determine if there was an effect on progression of the underlying disease.¹⁰⁵ Patients were randomized to 1 year of treatment with rofecoxib 25 mg daily, naproxen 220 mg twice daily, or placebo. One hundred twenty-two patients received rofecoxib, 118 naproxen, and 111 placebo. No specific handling of cardiovascular events was undertaken, and safety data were collected as part of the routine monitoring of the trial. Serious adverse events of stroke/transient ischemic attack were reported in 3 patients on rofecoxib, 3 on naproxen, and 1 patient on placebo. Serious adverse events of myocardial infarction (MI) were noted in 3 patients on rofecoxib, none on naproxen, and 1 patient on placebo. Because the number of serious adverse events was small, there was no suggestion of an increased risk of cardiovascular events with naproxen.¹⁰⁶

Graham and colleagues examined data from a large health insurance databank.¹⁰⁷ A total of 1,394,764 individuals aged 18 to 84 years received various NSAIDs during a 3-year interval, amounting to an observation period of 2,302,029 person-years. The investigators identified 8,143 cases of serious coronary heart disease, of which 2,210 were fatal. For naproxen versus non-naproxen NSAID use, the adjusted odds ratio was 1.14. The authors concluded that there was no significant difference between naproxen and control drugs with respect to their ability to protect against serious coronary heart disease. Naproxen neither increased the risk for cardiovascular events nor protected against them.

Juni and coworkers reviewed the published findings from 11 observational trials.¹⁰⁸ Their assessment was that the relative risk of MI associated with naproxen usage was 0.84 compared with both nonusers of NSAIDs and users of other NSAIDs. Naproxen use resulted in a small but significant cardioprotective effect. The results of their meta-analysis were later summarized in a commentary published in the Journal of the American Medical Association (JAMA) that concluded naproxen may be the only non-aspirin NSAID that possesses cardioprotective activity, with an estimated risk reduction of MI of 14% (compared to 23% reduction by aspirin).⁸¹

The finding of an apparent cardioprotective effect for naproxen in a subsequent (June 2006) publication supported the earlier meta-analysis of Juni et al. Kearney and coworkers published their results of a meta-analysis of 138 randomized clinical trials involving more than 145,000 patients.¹⁰⁹ They showed that:

• coxibs were associated with an increased relative risk of first serious vascular events by 42% compared with placebo, attributed primarily to a nearly twofold increase in MI;
• ibuprofen (800 mg 3 times daily) and diclofenac (75 mg 2 times daily) were both associated with an increased risk for cardiovascular events; and,
• naproxen (500 mg 2 times daily) did not show an increased risk of serious cardiovascular events.

The summary rate ratio for cardiovascular events, compared with placebo, was 1.51 for ibuprofen, 1.63 for diclofenac, and 0.92 for naproxen.

Two recent reports relative to NSAID risks may either clarify or confuse the issue further.¹¹⁰ In the first report, Zhang and colleagues addressed the issue of renal effects with coxibs.⁷⁴ They conducted a meta-analysis of 114 clinical trials involving 116,094 patients, in which rofecoxib, but not celecoxib, valdecoxib, or etoricoxib (an investigational new drug), was associated with increased risk of peripheral edema, hypertension, cardiac arrhythmias, and renal dysfunction. In the other report, McGettigan and Henry focused on the cardiovascular risk of coxibs and nonselective NSAIDs.¹¹¹ Their systematic review of 17 case-control studies (>600,000 patients) and 6 cohort studies (>1 million patients) revealed that the risk of cardiovascular events was increased with rofecoxib, diclofenac, indomethacin, and probably meloxicam. Naproxen neither increased nor decreased risk. Both reports, scheduled for publication in the October 4, 2006 issue of JAMA were made available a month ahead of schedule on the website of the American Medical Association.

Finally, the results of the Alzheimer’s Disease Anti-inflammatory Prevention Trial (ADAPT) were published in November 2006, 23 months after its early termination.¹⁸⁰ ADAPT, sponsored by the United States National Institute on Aging, started recruitment in 2001. The original plan specified that 2625 participants would be recruited to show whether use of celecoxib or naproxen over a 7-year period would lower the risk of developing Alzheimer’s disease (AD). Study investigators terminated the trial early when preliminary data implied that subjects taking either drug were at increased risk of cardiovascular events.

Examination of the published ADAPT results clearly show that the early data, which suggested increased risk of cardiovascular events, were flawed and that terminating the trial before its intended completion resulted in data that cannot be reliably interpreted.^180,181 The number of safety events were small. Moreover, the trial was designed primarily to examine prevention of AD; it was not designed or powered for cardiovascular or cerebrovascular events as primary outcomes.

Further clarification of the issue of cardiovascular safety was noted in a poster session at the 70^th Annual Meeting of the American College of Rheumatology, November 2006. The results of a randomized, placebo-controlled, crossover trial in healthy male and female subjects were reported.¹⁷³ Participants received naproxen sodium, 220 mg twice a day, 220 mg 3 times daily, 550 mg twice a day, or placebo for 7 days. Following a 6-day washout period, subjects were crossed over to receive enteric coated aspirin, 81 mg once daily, for 7 days. The primary endpoint was inhibition of serum TBX₂. The inhibitory effect of the OTC doses (220 mg 2 or 3 times daily) was similar to the Rx dose (550 mg twice daily). None of the doses were inferior to aspirin. These results further support an apparent cardioprotective effect for naproxen.

Questions regarding cardiovascular safety with NSAIDs remain unanswered. For patients with OA or other conditions who require relief of chronic pain, naproxen appears to be the safest NSAID choice.¹¹⁰ For patients at high risk of NSAID-related gastrointestinal events (Table 7), naproxen plus a proton-pump inhibitor^43,112,113 is less costly and as effective, and most likely safer, than celecoxib.^110,179

Table 7. Risk factors for development of upper gastrointestinal complications from NSAIDs¹⁷²

·         Prior ulcer and its complications

·         Age >75 years

·         Multiple high-dose NSAIDs

·         Use of warfarin

·         Use of alcohol

·         Significant comorbidity

Drug Interactions
NSAIDs may interact adversely with other drugs.¹¹⁴ Because they are highly protein-bound and compete with other drugs bound to the same protein binding sites, NSAIDs may displace orally administered anticoagulants, older hypoglycemic agents (tolazamide, tolbutamide), and anticonvulsants from their respective binding sites, thus increasing the clinical effects of these drugs. NSAIDs can reduce the effectiveness of antihypertensive agents, so blood pressure should be monitored when an NSAID is used concomitantly with an antihypertensive agent. NSAIDs inhibit lithium clearance; if this combination is required, careful monitoring of lithium levels is essential. Severe toxicity has been reported following administration of methotrexate at 15 mg or more per week with NSAIDs, due to decreased methotrexate excretion. Therefore, NSAIDs should be used with caution in patients receiving methotrexate therapy.¹¹⁵

NSAIDs plus low-dose aspirin. Millions of people in the United States take low-dose aspirin (eg, 81 mg/day) for cardioprotection.¹¹⁶ They face a potential problem when attempting to treat pain with NSAIDs: when used concurrently with aspirin, NSAIDs may increase the risk of gastrointestinal complications. Also, ibuprofen may interfere with the antithrombotic benefits of aspirin through competitive interaction with platelet COX-1.¹¹⁷

The FDA affirmed this interaction potential for ibuprofen in September 2006, when it notified healthcare professionals^118,119 and consumers and patients¹²⁰ about a potentially serious pharmacodynamic interaction that could pose serious adverse consequence. The notice advised that ibuprofen can attenuate aspirin’s effect when used for cardioprotection and stroke prevention. The agency informed healthcare professionals to counsel consumers and patients regarding the appropriate concomitant use of ibuprofen and aspirin. It advised consumers and patients to contact their physician for clarification on the timing of when to take these two drugs to maximize safety and efficacy (Table 8).

Table 8. FDA recommendations concerning concomitant use of ibuprofen and aspirin¹¹⁸

 ·       Healthcare providers should counsel patients about the appropriate timing of ibuprofen dosing if the patients
      are also taking aspirin for cardioprotective effects.

·         With occasional use of ibuprofen, there is likely to be minimal risk from any attenuation of the antiplatelet effect of low-dose aspirin.

·         Patients taking immediate-release low-dose aspirin (not enteric coated) and ibuprofen 400 mg should take the ibuprofen at least 30 minutes after aspirin ingestion, or at least 8 hours before aspirin ingestion to avoid any potential interaction.

·         Other nonselective OTC NSAIDs should be viewed as having potential to interfere with the antiplatelet effect of low-dose aspirin unless proven otherwise.

·         Analgesics that do not interfere with the antiplatelet effect of low-dose aspirin should be considered for populations at high risk for cardiovascular events.

·         Recommendations about concomitant use of ibuprofen and enteric-coated low-dose aspirin cannot be made based on available data.

The agency cited a report by Catella-Lawson and coworkers that demonstrated the potential for interaction.¹²¹ Healthy subjects received low-dose aspirin (81 mg; immediate-release tablets) in the morning either 2 hours before or 2 hours after ibuprofen (400 mg), rofecoxib (25 mg) or acetaminophen (1000 mg) for 6 consecutive days. Investigators assessed platelet COX activity by determining serum thromboxane B₂ (TXB₂) levels on day 7. Subjects who received ibuprofen 400 mg 2 hours before taking 81 mg of aspirin had serum TXB₂ inhibition of 53% on day 7, compared with 99% inhibition in those who took aspirin first. Platelet aggregation was irreversibly inhibited by 98% in patients who took aspirin first, versus inhibition of 2% in those who took ibuprofen first. Neither rofecoxib nor acetaminophen interfered with the effects of aspirin. Both ibuprofen and aspirin bind with COX, such that the presence of ibuprofen interferes with the binding of aspirin. This competition with aspirin’s binding to COX interferes with its irreversible inhibition of TXB₂ production and reduces its expected inhibition in platelet aggregation. Attenuation of 90% or more of the antiplatelet effect of aspirin has been defined as clinically significant.¹²²

In further investigation, Catella-Lawson et al administered enteric-coated aspirin at 8:00 a.m., followed by multiple doses of ibuprofen or diclofenac throughout the day—a schedule more in line with typical NSAID dosing by patients with OA. Group 1 took ibuprofen 400 mg 3 times daily beginning 2 hours after the first aspirin dose (10 a.m., 3 p.m., and 8 p.m.). Group 2 took diclofenac 75 mg 2 times daily (10 a.m. and 6 p.m.) for 6 days. Serum TXB₂ was measured 24 hours after administration of the first study drug. Levels were inhibited 67% in patients receiving ibuprofen compared with 92% when patients took diclofenac after aspirin. The investigators suggested that sufficient ibuprofen remained in the blood from the evening dose to interact with the next morning dose of aspirin.

Studies have subsequently confirmed that single doses of ibuprofen 400 mg, taken within 30 minutes after immediate-release aspirin dosing, interfere with aspirin’s antiplatelet activity, as measured by TXB₂ levels and platelet activation studies. The interaction also occurs with a single dose of ibuprofen 400 mg taken within 8 hours prior to aspirin dosing. Therefore, to avoid significant interaction, at least 8 hours should elapse following ibuprofen dosing before aspirin is given.¹¹⁸

Acetaminophen does not appear to interfere with the antiplatelet effect of low-dose aspirin.¹²¹ One study of naproxen, 500 mg twice daily, and low-dose aspirin, 100 mg daily, suggests that naproxen may interfere with aspirin’s antiplatelet activity when they are administered together. However, naproxen 500 mg administered 2 hours before or after the administration of aspirin 100 mg failed to interfere with aspirin’s antiplatelet effect.¹⁷⁴ The advantages and disadvantages of various NSAIDs should be weighed when deciding which agent is best for a particular patient (Table 9).¹²³

Other Therapies
Tramadol.—Tramadol is a synthetic codeine analogue with weak agonistic affinity for opioid receptors. It acts centrally to inhibit reuptake of norepinephrine and serotonin; this inhibition is responsible for at least part of its analgesic action. Tramadol is approved for treatment of moderate to severe pain and is a candidate for safe use in persons for whom NSAIDs are contraindicated or have failed to provide adequate pain relief.⁶ Few controlled clinical trials have examined tramadol for use in OA.

Adverse effects are common and include nausea and vomiting, constipation, dizziness, and respiratory depression.¹²⁴ Despite its chemical similarity to opioids, tramadol has failed to demonstrate significant abuse potential. It remains an unscheduled agent.¹²⁵

Opioids.—Patients with severe pain who cannot tolerate or do not respond to NSAIDs or tramadol may be considered for an opioid analgesic on an acute, short-term basis. Tolerance, dependence, and adverse effects, including respiratory depression and constipation, may occur with their use.¹²⁴

Disease-modifying antirheumatic drugs.—Disease-modifying antirheumatic drugs (DMARDs) (eg, gold salts, methotrexate, etc.) are used in selected arthritic and rheumatic diseases to prevent, slow, or reverse changes to the articular cartilage matrix. To date, there are no controlled clinical studies that show DMARDs are more effective than NSAIDs in treatment of OA.

Corticosteroids.—For the past 5 decades, corticosteroids, including triamcinolone, methylprednisolone, and prednisolone, have been injected intra-articularly to benefit OA patients who have local inflammation with joint effusion.^126-129 These agents may be considered after nonpharmacologic treatments and NSAIDs have failed.⁶ Intra-articular corticosteroid injections have demonstrated an excellent safety record and are described by the ACR as safe and effective when administered by an experienced physician.¹³⁰ As with any pharmacologic agent, corticosteroids can produce adverse effects and treatment should be given only for an appropriate indication. Most clinicians recommend at least a 3 month interval between injections, largely out of concern for the potential effects of corticosteroids on joint structure; however, data supporting this specific interval are lacking.¹³¹

The role of corticosteroid injections in the long-term management of OA remains unclear. Their use may be considered for short-term treatment or when the patient is otherwise unresponsive to both nonpharmacologic and other pharmacologic modalities.

Hyaluronic acid.—Injections of hyaluronic acid into the knee joint are approved for treatment of OA. Injections are generally well tolerated.¹³² However, data on efficacy are inconsistent. Two recent meta-analyses reported statistically significant but limited efficacy.^133,134 In one meta-analysis, publication bias (preferential publication of positive studies) was determined, which can inflate reported estimates.⁴³

External analgesics.—Externally applied analgesics offer modest benefit for some patients.^135,136 The remedies are used to self-treat mild-to-moderate arthritic pain localized to limited areas. Topical drug delivery is an optimal route for treatment of small areas because high drug concentrations can be localized at sites where needed, thereby minimizing the potential for systemic adverse effects, or interactions with food or other drugs.

Counterirritants (eg, methyl salicylate [oil of wintergreen], camphor, menthol, capsaicin) applied to the skin produce mild, local inflammatory reactions that help relieve pain indirectly by stimulating sensations of cold, warmth, or sometimes itching. Inducing mild pain or other neural sensation voluntarily as a means to counter a more intense one is instinctive. For example, pain sufferers may bite their lips, clench their fists, or massage or dig their nails into their skin in order to reduce perception of intense pain. Superficial sensations caused by counterirritants, therefore, distract from the deep-seated painful stimuli originating within inflamed or damaged joints. The intensity of response depends upon the specific counterirritant and its concentration, the vehicle used, and the extent of contact with the skin. Mere massaging of the skin during product application can help suppress pain transmission; thus, some benefit from counterirritants may involve additional components. There may also be a strong placebo component based upon many of the products’ characteristic “medicinal” odors.⁷⁶

Capsaicin, the major pungent ingredient obtained from the common hot chili pepper (Capsicum frutensens), desensitizes the vanilloid receptor, a gated cation channel.¹³⁷ Capsaicin also depletes substance P, a neuropeptide chemomediator in pain transmission, from small, type C unmyelinated nerve fibers.¹³⁸ Capsaicin must be applied three times daily, often for two or more weeks, for full benefit.

Patients using a product containing capsaicin or other rubefacient should be advised about the transient burning/stinging sensation that occurs after application. This localized adverse sensation normally decreases in intensity with continued use and rarely leads to discontinuation of therapy. Patients should be advised to massage the product thoroughly into the affected area. They should wash their hands carefully after applying the product and avoid contact between the product and the eyes and other mucous membranes.

A meta-analysis of 86 randomized controlled trials involving 10,160 patients relative to use of topical NSAIDs to treat a variety of conditions, including OA, concluded that the drugs effectively relieved pain of acute and chronic OA and had a low incidence of local and systemic adverse events.¹³⁹ The topically applied NSAIDs were superior to placebo in reducing pain; 65% of treated patients, versus 30% of controls, reported their pain was reduced by half. In another meta-analysis of randomized co