British Columbia Medical Journal, October 2010 issue Please download or visit this entire issue online at http://www.bcmj.org/october-2010

1. Stephen Kennedy, MD, Michael Moran, MBBS, FRCSC
Pharmacological treatment
of osteoarthritis of the hip
and knee
A range of oral analgesics, topical treatments, and intra-articular
injections can be used to reduce pain and improve function in
patients with osteoarthritis of the hip and knee.
reatment for hip and knee Evidence for the proposed mechanism
T
ABSTRACT: Treatment goals for the
management of hip and knee osteo- osteoarthritis (OA) aims to is insufficient in vivo, but some stud-
arthritis are to reduce pain, maintain reduce pain, maintain or im- ies have reported benefits from gluco-
or improve function, and, where pos- prove function, and, where samine in terms of pain relief and even
sible, to slow the progress of the possible, to slow the progress of the radiographic progression.1
underlying disease. The benefits and underlying disease. Although no med- Evidence for a positive effect is
potential toxicities of pharmacologi- ication has yet been shown to slow the controversial, however, with several
cal options should be considered advance of joint pathology in osteo- studies showing no benefit over pla-
and treatment should be individ- arthritis, pharmacological manage- cebo.1,2 The Osteoarthritis Research
ualized according to patient symp- ment remains an integral component Society International (OARSI) guide-
toms, preferences, and a therapeu- of therapy for most patients in the lines state that “treatment with glu-
tic agent’s overall safety profile. course of their disease. cosamine and/or chondroitin sulphate
When intrusive pain or disability Knowledge of the current evi- may provide symptomatic benefit in
persists despite a substantial trial dence can support safe and effective patients with knee OA,” but “if no
of nonsurgical therapy, or when con- counseling and prescribing practices, response is apparent within 6 months
trolling symptoms requires long- and assist in determining when to treatment should be discontinued.”2
term opioids, high-dose acetamin- refer for surgery. As well as consid- Other guidelines, such as those for the
ophen or NSAIDs, and repeated ering the benefits and potential toxi- American Academy of Orthopaedic
intra-articular injections, referral cities of pharmacological options, Surgeons, make a recommendation
for surgical options should be con- physicians must individualize treat- that physicians not prescribe glu-
sidered. ment based on patient symptoms and cosamine.3 Both sets of guidelines are
preferences. based on level I evidence.2,3
This disparity in recommend -
Glucosamine sulfate and ations is considered to be due to the
chondroitin sulfate heterogeneity of existing studies,
Glucosamine is one of the most com- particularly with respect to adequacy
monly used complementary or alter-
native medicine products in North Dr Kennedy is a resident in orthopaedics at
America. Typically derived from the the University of British Columbia. Dr
ground shells of shellfish or from Moran is a clinical assistant professor and
processed grains, glucosamine has head of the Division of Comprehensive
proponents who claim it restores gly- Orthopaedics in the Department of
cosaminoglycans in arthritic joints Orthopaedics at the University of British
and reduces pain and inflammation.1 Columbia in Prince George, BC.
404 BC MEDICAL JOURNAL VOL. 52 NO. 8, OCTOBER 2010 www.bcmj.org

2. Pharmacological treatment of osteoarthritis of the hip and knee
of allocation concealment.2 Little or Higher doses of acetaminophen or sidered.2,3 Gastroprotection is recom-
no benefit has been observed when even prolonged use at recommended mended in all eight of the guidelines
concealment is adequate.2 Evidence doses are not without risk.2 Although where NSAIDs are considered for the
regarding chondroitin sulfate is sim- not common in the studies referenced management of hip or knee OA.2
ilarly inconsistent.2 There is marked by the guidelines above, acetaminophen COX-2 inhibitors are recommended
heterogeneity of outcomes between overdose can result in hepatoxicity in all 11 of the guidelines where they
trials, and again higher-quality stud- and severe sequelae. Patients should are considered.2 H2-receptor antago-
ies with adequate concealment have be counseled and monitored regard- nists do not have similar protective
been unable to show significant ben- ing their daily dosage. In the absence qualities, and the GI benefit associat-
efit.1,2 of an adequate response, or in the pres- ed with COX-2 agents is lost with con-
Overall, the evidence and recom- ence of severe pain or inflammation current low-dose daily acetylsalicylic
mendations remain inconsistent for (or both), alternative therapy should acid.2,3
both glucosamine and chondroitin.1-3 be considered. Combining acetamin- Cardiovascular (CV) risk is anoth-
We do not recommend prescription of ophen with another medication (e.g., er concern. After rofecoxib was with-
these supplements as their benefit ibuprofen) at lower doses of each can drawn from the market due to in-
remains unproven, but the risk of their also be effective. creased risk of thrombotic events, a
use seems limited to mild stomach number of studies were done to inves-
upset and the cost of the pills.1-3 A trial NSAIDs tigate the CV safety of other NSAIDs.
of treatment for 6 months would not NSAIDs or nonsteroidal anti-inflam- Celecoxib and valdecoxib do not
be unreasonable if a patient expresses matory drugs are among the most appear to have the same risks, and
great interest in such products. Future commonly used analgesics in the overall CV risk with COX-2 inhibitors
independent high-quality studies are world and are often used as first-line has not been found significantly high-
required to further clarify the efficacy medications for joint pain. One UK er than with nonselective NSAIDs.2
of both agents. telephone survey in 2003 reported that Serious vascular events occur at ap-
50% of respondents with osteoarthri- proximately 1% per year on COX-2
Acetaminophen tis were taking NSAIDs.2 inhibitors versus 0.9% on traditional
Acetaminophen is a common first-line There is good level I evidence for NSAIDs.2
analgesic for treatment of hip and the analgesic effect of NSAIDs in CV risk is greater in patients with
knee osteoarthritis. OARSI found the OA, and meta-analyses of short-term, a history of ischemic heart disease or
use of acetaminophen to be a core rec- placebo-controlled randomized trials stroke, or in patients with risk factors
ommendation in 16 of 16 guidelines have shown an effect size between 0.23 for heart disease such as hypertension,
evaluated.2 Compared with placebo, and 0.32 in terms of reduction in pain.2 hyperlipidemia, diabetes, smoking, or
statistically significant effects on pain NSAIDs, however, are associated peripheral arterial disease.2 Caution
relief have been demonstrated with- with more adverse effects than acet- should be exercised when prescribing
out statistically significant risk of tox- aminophen.2,3 Gastrointestinal (GI) all NSAIDs in these patients.2
icity.3 OARSI guidelines recommend discomfort occurs more frequently Renal toxicity is also a concern in
up to 4 g per day as an effective first- and, more importantly, serious com- selected patients. In patients with con-
line therapy in patients with mild to plications such as peptic ulcers, perfo- gestive heart failure, pre-existing
moderate pain from OA.2 Current rations, and bleeds are more likely to renal insufficiency, or transplanted
European League Against Rheuma- occur.2,3 Pooled relative risk compared kidneys, the use of NSAIDs can lead
tism (EULAR) recommendations for to placebo is estimated at 270%.2 Risk to acute renal failure. Care should be
hip and knee OA suggest that aceta- also increases with age, concurrent taken to screen for clinical or labora-
minophen at these doses should be the use of other medications, and duration tory evidence of existing diminished
first choice for mild to moderate pain, of therapy. creatinine clearance and considera-
and if successful, should be used as In patients at greater GI risk, there tion should be given to follow-up lab
the preferred long-term oral anal- is level I evidence that NSAIDs should analysis after treatment is begun.
gesic.2 For most patients the differ- be used in combination with a proton Renal clearance decreases significant-
ence in pain relief between acetamin- pump inhibitor or misoprostol for ly with age.
ophen and NSAIDs is not clinically gastroprotection, or that the use of a In patients with symptomatic hip
significant.2 COX-2 selective agent should be con- or knee OA, NSAIDs should be used
www.bcmj.org VOL. 52 NO. 8, OCTOBER 2010 BC MEDICAL JOURNAL 405

3. Pharmacological treatment of osteoarthritis of the hip and knee
at the lowest effective dose and their studies. This compared with 7% of sizes of 0.41 and 0.40, respectively, in
long-term use should be avoided if placebo-treated patients.2 weeks 1 and 2.2 Side effects seem lim-
possible.2,3 In patients at greater GI There have been no long-term tri- ited to local reactions such as burning,
risk, either a COX-2 selective agent or als of the use of opioids for OA, and itching, and rashes.2 Placebo effects
a nonselective NSAID in combination ongoing concerns remain about the may be large with topical therapies,
with a gastroprotective agent should risks of dependence. Recovery from and one meta-analysis showed evi-
be considered.2,3 All NSAIDs should arthroplasty surgery is more difficult dence of possible publication bias
be used with caution in patients with for patients on chronic opioid therapy, with underreporting of negative stud-
CV risk factors.2 Physicians should and their optimal outcome may be ies.2 However, topical NSAIDs re-
continue to choose an NSAID on the compromised.4 main a reasonable option in combina-
basis of the agent’s overall safety pro- We feel that strong opioid anal- tion with or as an alternative to other
file and the patient’s individual risk gesics should be reserved for patients analgesics.
factors. in exceptional circumstances with
severe pain who are not candidates for Topical capsaicin
Opioids other therapy. Short courses of weak Topical capsaicin creams contain a
Weak opioids have increasingly been opioids like codeine or tramadol and lipophilic alkaloid extracted from
used recently for the treatment of acetaminophen combinations can be chili peppers that activates and sensi-
refractory pain in patients with hip or used for brief exacerbations of pain if tizes peripheral pain and heat recep-
knee OA. A number of systematic tolerated.2 When prescribing these, tors by binding and activating specif-
reviews and meta-analyses of opioids precautions should be taken: patients ic cation channels.2 Application to the
for chronic non-cancer pain, musculo- should be counseled about their use skin causes a burning sensation ini-
skeletal pain, and OA have provided and potential for dependence. Non- tially but can lead to effective analge-
evidence of efficacy and acceptable pharmacological therapies should con- sia that prevails over the sensation of
safety in short-term trials.2 tinue and surgical treatments should burning.2 The efficacy of capsaicin is
Analysis of 18 randomized place- be considered. It is highly recom- supported by a meta-analysis of RCTs
bo-controlled trials of 3244 OA pa- mended that strong narcotics such as of its use in the treatment of chronic
tients showed a moderate effect size morphine, oxycodone, and hydromor- painful conditions, including a single
for reduction in pain intensity (0.25).2 phone not be prescribed for osteo- placebo-controlled trial in 70 patients
However, there was substantial het- arthritis. Instead, patients should be with knee OA and two RCTs in
erogeneity between studies. This was referred for surgical treatment. patients with hand OA.2 The mean
not obviously related to the prepara- reduction in pain was 33% after 4
tion used or the quality of the RCTs.2 Topical treatments weeks of therapy.2 Treatment is safe,
A systematic review regarding Topical NSAIDs but local burning, stinging, or erythe-
acetaminophen and codeine combina- Topical NSAIDs can be effective ma troubles 40% of patients. The
tions indicated a small analgesic ben- adjunctive treatments or alternatives burning sensation also prevents ade-
efit over acetaminophen alone (approx- to oral analgesics in knee OA.2 A quate blinding with this agent, which
imately 5%), but adverse effects were meta-analysis of 13 RCTs, including may influence conclusions based on
more frequent.2 Another meta-analysis 1983 patients with hand and knee OA, the available data.2 Despite these
of opioids for chronic non-cancer showed topical NSAIDs to be superi- shortcomings, topical capsaicin can
pain, including OA, demonstrated that or to placebo in terms of analgesia, be a useful alternative or adjunctive
only strong opioids were significantly relief of stiffness, and function, with a treatment in selected patients.2 A typ-
more effective in relieving pain than reduced relative risk of adverse GI ical dose is 0.025% cream four times
acetaminophen or NSAIDs.2 events compared with oral forms.2 In a day.2
Benefits associated with the use of one large case control study topical
opioids, however, are limited by fre- NSAIDs were reported to have no Intra-articular injection
quent side effects such as nausea more GI side effects than placebo.2 Techniques for injection
(30%), constipation (23%), dizziness Topical NSAIDs are less effective Intra-articular injection of the hip gen-
(20%), somnolence (18%), and vom- than oral NSAIDs in the first week of erally requires fluoroscopic or ultra-
iting (13%).2 One-quarter of patients treatment, but efficacy is apparent sound guidance to ensure accurate
treated with opioids withdrew from within 2 weeks, with pain relief effect placement.
406 BC MEDICAL JOURNAL VOL. 52 NO. 8, OCTOBER 2010 www.bcmj.org

4. Pharmacological treatment of osteoarthritis of the hip and knee
Multiple descriptions exist for cerns in terms of “trial quality, poten- Discordant conclusions in system-
intra-articular injection of the knee tial publication bias, and unclear clin- atic reviews of HA have been found
joint.5,6 The patient should be supine ical significance.”2 Pooled effects to be due to inclusion of different
and relaxed. It is easiest to inject the from poor-quality trials are as much controlled trials, differences in the
knee in full extension. All injections as twice those obtained from higher- outcome measures and time points
should be performed in a sterile man- quality ones.2 In systematic reviews selected for extraction, and different
ner.5,6 It’s helpful to palpate surface
landmarks prior to antiseptic cleans-
ing and draping. A 25-gauge 11/2 inch
needle should be used. One study found
the lateral mid-patellar approach to Injection of hyaluronic acid
have the greatest accuracy.5 More
common, and our method of choice, is preparations into the knee and hip is
to use the soft point at the superior lat- commonly used to treat osteoarthritis, but
eral pole of the patella between the
patella and the femur, with the needle there is considerable ongoing controversy
inserted into the suprapatellar pouch about the treatment’s efficacy, cost-
at that level. Entry should be deliber-
ate and smooth. Joint effusion can effectiveness, and benefit-to-risk ratio.
make the process much easier, while
factors such as joint degeneration,
diminished range of motion, and obe-
sity can make insertion more diffi- there is significant heterogeneity be- statistical methods for data synthesis,
cult.5,6 If the needle meets an obstruc- tween studies and evidence to suggest which resulted in conflicting esti-
tion, pull back slightly and adjust the publication bias and overestimation mates of therapeutic effect.2
trajectory. Aspiration of joint fluid can of effect size.2 No major safety issues were de-
be used for confirmation of accurate A Cochrane review of 40 placebo- tected, but in placebo-controlled trials
placement. During injection, patient controlled trials with five different minor adverse events such as transient
complaints of increased pain should hyaluronan products found statistical- pain at the injection site occurred
be considered an indication of possi- ly significant improvements in pain slightly more frequently in patients
ble extra-articular placement.5,6 The on weight bearing when results were treated with intra-articular hyaluro-
fluid should flow smoothly and cause pooled, but improvements were vari- nan than in those treated with intra-
little or no discomfort. If infiltration is able.2 Pain reduction from baseline at articular corticosteroids.2
difficult, reposition and reattempt 5 to 13 weeks varied from 28% to 54% Because of the conflicting evidence
injection as necessary. for pain and 9% to 32% for function- from the literature and existing guide-
al outcome scores.2 Data to suggest lines, the use of intra-articular HA
Viscosupplementation that the higher molecular weight HA is not universally recommended.2,3
Hyaluronic acid (HA) or hyaluronan preparations were more effective than Relief may be gained for patients with
is a glycosaminoglycan constituent of lower molecular weight preparations mild to moderate hip or knee OA
synovial fluid. Injection of HA prepa- were inconclusive.2 In a randomized symptoms, and results are character-
rations into the knee and hip is com- comparison of three injections of high ized by delayed onset but prolonged
monly used to treat osteoarthritis, but and low molecular weight HA, there duration.2 The adequacy of clinical
there is considerable ongoing contro- were significant improvements of benefit remains somewhat unclear and
versy about the treatment’s efficacy, approximately 40% in pain and func- costs are not insignificant—injections
cost-effectiveness, and benefit-to-risk tional scores up to 6 months after treat- typically range from $130 to $230 per
ratio.2,3 ment.2 However, in another placebo- injection and 3 to 5 weekly injections
Numerous studies have examined controlled trial comparing HA with are required. We tend not to recom-
the effectiveness of various HA prepa- corticosteroid or saline at 2 weekly mend these injections, particularly in
rations and generally show positive intervals, there were no significant patients with moderate to severe dis-
effects, but there are significant con- differences between the groups.2 ease, but if patients are given realistic
www.bcmj.org VOL. 52 NO. 8, OCTOBER 2010 BC MEDICAL JOURNAL 407

5. Pharmacological treatment of osteoarthritis of the hip and knee
expectations and have adequate re- may be due to additional mechanisms to-head comparisons exist to support
sources, a trial of therapy is not unrea- unrelated to the purely anti-inflamma- any particular choice of corticosteroid,
sonable, particularly for mild OA. tory effect. and data are insufficient to state how
In terms of toxicity, potential side frequently it is safe to repeat injec-
Corticosteroid therapy effects include post-injection flares of tions. More than four times annually
Despite the unclear role of inflamma- pain, crystal synovitis, hemarthrosis, is generally not recommended. One
tion in the pathogenesis and progres- joint sepsis, articular cartilage atro- indication for these injections is if a
sion of osteoarthritis, 11 of 13 existing phy, and steroid-induced arthropathy. patient needs to be active for a short
treatment guidelines recommend in- Side effects such as bruising and period of time while awaiting surgery,
jection of corticosteroids for OA at lipodystrophy are not uncommon but either because of work or family com-
some stage of the disease.2 Multiple can be minimized with careful tech- mitments. The temporary relief, partic-
systematic reviews conclude that it is nique. Overall, in 28 controlled trials ularly if the patient is clearly informed
effective for relieving pain at least in of intra-articular steroid injections in about its temporary nature, is often
the short term (i.e., 1 to 2 weeks).2,3,7 1973 patients with OA of the knee, no appreciated.
The efficacy is also supported by evi- serious adverse events were reported
dence from a Cochrane systematic as a consequence.2 In cases where Antidepressants
review, which examined data from 13 inflammatory or infectious arthritis is Depression and osteoarthritis are both
randomized placebo-controlled tri- considered, aspiration and analysis of common and often coexist. Multiple
als.2,8 The effect size for pain relief is synovial fluid prior to injection should studies have demonstrated that psy-
in the moderate range (0.25) at 2 and also be considered. chosocial factors are equally or more
3 weeks after injection, with a lack of OARSI guidelines state that intra- important than disease-specific fac-
evidence for pain relief by 4 weeks articular injections with corticosteroid tors in reports of pain intensity and
and 24 weeks after injection.2,8 Evi- can provide short-term symptomatic disability in several conditions, in-
dence for hip steroid injection is more relief of knee OA, and should be con- cluding joint pain.9,10 Awareness and
limited, and mixed in terms of results.2 sidered, particularly in cases of mod- treatment of depressive symptoms can
result in significantly less pain and
improved quality of life.9,10 In one
study of older adults with arthritis and
comorbid depression, treatment of
Despite the unclear role of inflammation depression extended beyond improved
mood to significant improvement in
in the pathogenesis and progression of pain, function, and quality of life.10
osteoarthritis, 11 of 13 existing treatment
When to refer
guidelines recommend injection of The decision to refer a patient for sur-
corticosteroids for OA at some gery is complex, and consensus state-
ments fail to agree upon specific
stage of the disease. thresholds for referral, but pain and
disability are consistently the most
important measures considered.11-13
Physical examination is emphasized
less and tends not to correlate with the
Some randomized controlled trials erate to severe pain not responding decision for surgery.12,13 The ability to
have demonstrated better outcomes in to other analgesics and nonpharma- work, give care to dependants, and
patients with synovial effusions or cologic modalities.2 Anecdotally we live independently consistently out-
other clinical signs of inflammation, have found a small percentage of pa- weigh range of motion or other meas-
but this has not been seen universally tients to achieve long-term improve- ures of physical impairment.13
and it remains controversial whether ment. For the most part, however, Generally, referral to an orthopaedic
steroid injections should be restricted improvements are short-lived for what surgeon should be made when intru-
to these patients.2 The analgesic effect is a chronic problem.7 Too few head- sive pain or disability persists despite
408 BC MEDICAL JOURNAL VOL. 52 NO. 8, OCTOBER 2010 www.bcmj.org

6. Pharmacological treatment of osteoarthritis of the hip and knee
a substantial trial of nonsurgical ther- Competing interests 11. Naylor CD, Williams JI. Primary hip and
apy, or when long-term opioids, high- None declared. knee replacement surgery: Ontario crite-
dose acetaminophen or NSAIDs, and ria for case selection and surgical priority.
repeated intra-articular injections are References Qual Health Care 1996;5:20-30.
required to control symptoms. The 1. Block JA, Oegema TR, Sandy JD, et al. 12. Hadorn DC, Holmes AC. The New
decision should be personalized and The effects of oral glucosamine on joint Zealand priority criteria project. Part 1:
based on each patient’s experience of health: Is a change in research approach Overview. BMJ 1997;314:131-134.
the disease, functional goals, and risks needed? Osteoarthritis Cartilage 2010; 13. Arnett G, Hadorn DC. Developing priori-
of undergoing elective surgery. 18:5-11. ty criteria for hip and knee replacement:
Age, obesity, and comorbidities 2. Zhang W, Moskowitz RW, Nuki G, et al. Results from the Western Canada Wait-
have little impact on the benefit from OARSI recommendations for the man- ing List Project. Can J Surg 2003;46:290-
joint replacement and rarely should agement of hip and knee osteoarthritis, 296.
prevent referral.14 Hip and knee replace- part II: OARSI evidence-based, expert 14. Santaguida PL, Hawker GA, Hudak PL.
ment surgeries reliably reduce pain, consensus guidelines. Osteoarthritis Patient characteristics affecting the prog-
restore function, and have low mor- Cartilage 2008;16:137-162. nosis of total hip and knee joint arthro-
bidity and mortality.15,16 With improv- 3. Richmond J, Hunter D, Irrgang J, et al. plasty: A systematic review. Can J Surg
ing joint arthroplasty survivorship, it Treatment of osteoarthritis of the knee 2008;51:428-436.
has also become a viable option for (nonarthroplasty). J Am Acad Orthop 15. Schulte KR, Callaghan JJ, Kelley SS, et al.
younger patients with disabling dis- Surg 2009;17:591-600. The outcome of Charnley total hip arthro-
ease. Decision making in these cases 4. Carroll IR, Angst MS, Clark JD. Manage- plasty with cement after a minimum
can be challenging, and orthopaedic ment of perioperative pain in patients twenty-year follow-up. The results of one
consultation may be the best way to chronically consuming opioids. Reg surgeon. J Bone Joint Surg Am 1993;
determine suitability for surgery. It is Anesth Pain Med 2004;29:576-591. 75:961-975.
important to note that there is no age 5. Jackson DW, Evans NA, Thomas BM. 16. Robertsson O, Dunbar M, Pehrsson T, et
restriction for joint replacement sur- Accuracy of needle placement into the al. Patient satisfaction after knee arthro-
gery, and referral should not be with- intra-articular space of the knee. J Bone plasty: A report on 27,372 knees operat-
held because of young age. Joint Surg Am 2002;84:1522-1527. ed on between 1981 and 1995 in
6. Lockman LE. Knee joint injections and Sweden. Acta Orthop Scand 2000;71:
Conclusions aspirations: The triangle technique. Can 262-267.
Pharmacological treatments for osteo- Fam Physician 2006;52:1403-1404.
arthritis of the hip and knee have not 7. Hepper CT, Halvorson JJ, Duncan ST, et
been shown to alter the progression of al. The efficacy and duration of intra-artic-
disease but may be used in a multitude ular corticosteroid injection for knee
of combinations for symptom relief. A osteoarthritis: A systematic review of
range of oral analgesics, topical treat- level 1 studies. J Am Acad Orthop Surg
ments, and intra-articular injections 2009;17:638-646.
of hyaluronic acid or steroids might 8. Bellamy N, Campbell J, Robinson V, et al.
be considered. Treatment should be Intra-articular corticosteroid for treat-
individualized according to patient ment of osteoarthritis of the knee.
symptoms, preferences, and a thera- Cochrane Database Syst Rev 2006;(2):
peutic agent’s overall safety profile. CD005328.
When intrusive pain or disability per- 9. Vranceanu AM, Barsky A, Ring D. Psy-
sists despite a substantial trial of non- chosocial aspects of disabling muscu-
surgical therapy, or when controlling loskeletal pain. J Bone Joint Surg Am
symptoms requires long-term opioids, 2009;91:2014-2018.
high-dose acetaminophen or NSAIDs, 10. Lin EH, Katon W, Von Korff M. Effect of
and repeated intra-articular injections, improving depression care on pain and
referral for surgical options should be functional outcomes among older adults
considered. with arthritis: A randomized controlled
trial. JAMA 2003;290:2428-2434.
www.bcmj.org VOL. 52 NO. 8, OCTOBER 2010 BC MEDICAL JOURNAL 409