➢ Side effects of corticosteroids can occur in association with both intra-articular and extra-articular injections.
➢ Less-severe reactions include skin reactions and flare responses.
➢ Although rare, worrisome complications such as infections, tendon ruptures, and osteonecrosis can occur.
➢ It is important to note the effect of corticosteroids on blood glucose levels, particularly for patients with diabetes.
➢ Corticosteroid injections have visual side effects with a known relation to central serous chorioretinopathy.
Injectable corticosteroids were first used in the knee in 19501. Locally, corticosteroid injections reduce inflammation in synovial tissues. This process of reduction is mediated by inhibition of phospholipase A22, reducing the presence of leukotrienes, prostaglandins, and thromboxanes, which are part of the inflammatory pathway. There is also a reduction in the number of lymphocytes, macrophages, and mast cells, which work in combination to reduce edema1. New formulations of corticosteroids are complexed with salts or polymers and are suspended in aqueous solutions to sequester the drug from the synovial fluid, delaying clearance from the joint space3. There is an inverse relationship between water solubility and length of action; more-insoluble corticosteroids remain at the site of injection for longer, possibly increasing the benefit4. Current commonly used corticosteroids include betamethasone sodium phosphate (the most soluble), methylprednisolone acetate, and triamcinolone (the least soluble)2. A survey found that physicians most often use the latter 2 for shoulder injections and, interestingly, do not vary the corticosteroid for different conditions5. However, some authors have recommended a more-soluble formulation for acute inflammatory conditions and a less-soluble formulation for chronic inflammatory conditions2.
Corticosteroids have been used for both intra-articular and extra-articular applications to reduce pain associated with osteoarthritis, rheumatoid arthritis, acute traumatic arthritis, tendinitis, bursitis, and ligament sprains1. Although the effect is mostly limited to short-term pain reduction, corticosteroids have been used successfully for the treatment of de Quervain tenosynovitis and trigger finger1,6,7. According to the American Academy of Orthopaedic Surgeons (AAOS) Clinical Practice Guidelines, there is strong evidence for the use of corticosteroid injections for the treatment of carpal tunnel syndrome but there is insufficient evidence to recommend for or against their use for the treatment of rotator cuff tears, glenohumeral arthritis, or knee-joint arthritis8-11.
Injections can be given in the outpatient setting, which is useful in that it allows the physician to obtain immediate feedback on the efficacy of the injection and to help diagnose a disease process when local anesthetic is used with the corticosteroid. Performing injections in the outpatient setting can add to the perception that the procedure is benign; however, complications can still arise. Some are well known, including septic arthritis, corticosteroid flare reactions, increased blood glucose levels, and tendon ruptures, whereas others are less familiar, such as central serous chorioretinopathy, tibial osteonecrosis, chondrotoxicity, and skin changes. Therefore, the purpose of the present review is to examine these potential complications and any related risk factors.
It generally has been recommended that joints should not receive more than 3 to 4 injections per 12 months3,12,13; however, others have reported that an injection can be repeated after 6 weeks4. Injections tend to be used initially before surgery is recommended. In the study by Sears et al., 84% of 18,335 patients with de Quervain tenosynovitis were initially treated with injection and 33% of those patients were given at least a second injection before surgery, 29% of 129,917 patients with carpal tunnel syndrome were initially treated with injection and one-third of those patients were treated with additional injections, and 74% of 102,778 patients with trigger finger were treated initially with a corticosteroid injection and 42% of those patients received >1 injection14. Kerrigan and Stanwix15 reported that 2 corticosteroid injections before surgery is the most cost-effective strategy for the treatment of trigger finger.
Absolute contraindications to local corticosteroid injection include local infection, bacteremia, fracture, joint prosthesis, tumor, Achilles or patellar tendinopathy, and history of allergy to either the corticosteroids or constituents within the injectable preparation. Relative contraindications should be determined on a case-by-case basis and should include anticoagulation, uncontrolled diabetes, and minimal relief after 2 previous injections at the same site.
One of the most commonly recognized, and harmful, complications of corticosteroid injections is septic arthritis. The reported incidence varies from 1:1,000 to 1:16,00016,17. This range is consistent with the perceptions of physicians, with 48% and 32% of 853 physicians perceiving the risk of infection to be 1:1,000 and 1:10,000, respectively18. The causative bacteria have included Staphylococcus aureus, S. epidermidis, and Escherichia coli17,19. Von Essen and Savolainen16 proposed 4 mechanisms for the etiology of post-injection septic arthritis: (1) contamination of the injected drug preparation, (2) introduction of bacteria from the skin, (3) hematogenous colonization of the puncture track, or (4) soft-tissue trauma from the injection, resulting in a hormonal activation of a previous quiescent infection. In a series of 223 cases of infection following injection that were reviewed by medical expert committees and mediation boards, 34% of infections were found to have been due to treatment error, usually resulting from faulty aseptic technique and rarely resulting from a faulty injection20. In a case series of 6 patients with septic arthritis after intra-articular injection, 50% had a corticosteroid injection, further suggesting that technique is more important than what is being injected as the cause of septic arthritis21. In the study by Charalambous et al., questionnaires regarding antiseptic technique during corticosteroid injection of the knee were sent to 250 orthopaedists, rheumatologists, and general practitioners22. All 191 respondents used some antiseptic (with 57.6% using alcohol alone and 42.4% using Betadine [povidone-iodine] or chlorhexidine), 82.6% did not use towels, 53.4% wore gloves, and 91.1% changed needles between drawing and injecting. While 12.6% of respondents had experienced at least 1 patient with septic arthritis, no correlation was drawn between technique and infection. Other treatment errors associated with post-injection infections include lack of documentation and failure to inform patients of risks and alternative treatments20.
Rheumatoid arthritis is another risk factor for the development of septic arthritis, particularly as the disease burden increases. Ostensson and Geborek reported that patients with rheumatoid arthritis who developed septic arthritis had a worse functional level, an increased number of intra-articular injections, and increased use of systemic cytotoxic drugs in comparison with a matched control group of patients with rheumatoid arthritis but without septic arthritis17. Munigangaiah et al. reported a rare instance of bilateral septic arthritis in a 66-year-old man with rheumatoid arthritis19. The patient also had end-stage renal disease and was undergoing hemodialysis, highlighting the extent of the overall disease burden, and he developed symptoms of infection after bilateral intra-articular corticosteroid injection involving the knee.
There is also concern that corticosteroid injections increase the risk of infection for patients who will soon have a total knee arthroplasty (TKA), particularly because as many as 30% of patients receive at least 1 corticosteroid injection prior to TKA23. Marsland et al. reported that 3 of the 4 studies that they reviewed demonstrated no increase in the infection rate23. Injections were often done at least a year before TKA; however, some were done within a month before surgery. Only 1 study demonstrated a possible association between injection and post-TKA infection when the injection was done between 8 and 11 months before the knee replacement23. A recent meta-analysis of 6 studies24 demonstrated no relationship between corticosteroid injection and deep or superficial infection.
If septic arthritis occurs, early diagnosis and treatment is important. Presenting symptoms may include erythema, effusion, fever, painful range of motion, and difficulty weight-bearing. Inflammatory laboratory markers and the white blood-cell count also will be elevated. Most patients present with symptoms within 1 to 3 days17,21 and require intravenous antibiotics and at least 1 irrigation-and-debridement procedure17,19,21.
In patients receiving extra-articular injections, the surrounding structures can still become infected. Woon et al. described the case of a 55-year-old man with type-2 diabetes who was given corticosteroid injections in the dorsal second through fifth metacarpophalangeal (MCPJ) joints twice over a 10-day period25. Two months later, the patient developed pain, swelling, and loss of active MCP joint extension. Intraoperatively, he was found to have suppurative ruptures of the extensor indicis proprius and extensor digitorum communis tendons of the second through fourth fingers. Cultures were positive for methicillin-resistant S. aureus, and the patient was eventually able to undergo reconstruction with use of a palmaris longus tendon graft. Our review of the literature revealed no reports of increased infection rates based on injection location.
Corticosteroid Flare Reaction
Although not nearly as dangerous as septic arthritis, a corticosteroid flare reaction is more common and can cause increased discomfort and symptoms. No exact etiology for the flare response has been determined, but it has been proposed that either the corticosteroid crystals themselves26 or the rapid intracellular ingestion of microcrystalline corticosteroid ester27 incite the reaction, and the reaction can be seen in association with variable formulations of corticosteroids. Studies ranging in size from 11 to 1,754 patients have shown the occurrence to be between 2% and 50%7,28-32, and corticosteroid flare reaction is frequently noted to be the most common side effect of corticosteroid injection7,30. In the study by Goldfarb et al.33, patients with trigger finger or de Quervain tenosynovitis were managed with either an injection of a combination of methylprednisolone, lidocaine, and bupivacaine (57 patients) or a balanced pH injection that added bicarbonate (68 patients). There was no difference between the 2 groups in terms of the rate of occurrence of a flare reaction (31.6% and 33.8%, respectively). Additionally, Price et al. reported no difference in flare response between 2 formulations of corticosteroids (hydrocortisone and triamcinolone) or between 2 doses of triamcinolone29. Future research could continue to delineate any difference in the rate of corticosteroid flare between different formulations.
Differentiating between a corticosteroid flare and septic arthritis can be challenging. Typically, corticosteroid flares occur sooner than septic arthritis following an injection and dissipate. A flare usually occurs during the first or second day30,33. Berger and Yount27 reported on a rare case of early-onset corticosteroid flare in a 32-year-old man with right knee pain who received an intra-articular injection and returned 90 minutes later with severe pain. A second injection did not alleviate the pain, which improved only after 24 hours of treatment with anti-inflammatory medications and bed rest. In terms of length, flare reactions can dissipate starting after an hour29 but have been reported to last up to 4, 7, and even 10 days29,31,33. There appears to be no relationship between the occurrence of a flare reaction and effectiveness of the corticosteroid at 6 weeks33. Treatment of corticosteroid flare includes rest, ice, and anti-inflammatory drugs.
Extra-articular corticosteroid injections carry a specific risk of tendon ruptures. In the study by Wong et al., human tenocytes from patellar tendons that were treated with dexamethasone had a loss of the normal bipolar spindle shape, with the tenocytes becoming more flattened and polygonal34. Furthermore, there was a significant decrease in the number of viable tenocytes (p < 0.001), decreased cell proliferation (p = 0.0063)—both in a dose-dependent manner—and decreased collagen production (p < 0.05). Interestingly, the addition of platelet-derived growth factor reversed all of these effects.
However, the research in vivo has mostly been reported through case studies. Smith et al. reported on a 45-year-old woman with lateral elbow tendinitis that was treated with injection of triamcinolone35. The elbow became swollen and painful, with notable pallor of the skin and a large defect at the extensor origin. At the time of surgery, rupture of 75% of the origin of the common extensor tendon was found. Fitzgerald et al. reported the case of a 77-year-old man who was given 2 injections of triamcinolone, 16 months apart, for the treatment of trigger finger36. After the second injection, the patient reported mild pain and difficulty making a fist. Eleven months later, he was unable to flex the distal interphalangeal joint or proximal interphalangeal joint. At the time of surgery, rupture of the flexor digitorum superficialis and profundus tendons was discovered. Histological examination of the tendons showed acute and chronic necrotizing tenosynovitis with a focal foreign-body multinucleated giant-cell reaction. Gyuricza et al. also reported on trigger finger, although with the occurrence of pulley rupture37. In that case, a 42-year-old woman had received 3 doses of triamcinolone within 12 months; however, she continued to have a flexion contracture and required release of the A1 pulley. Intraoperatively, absence of the proximal aspect of the A2 pulley was noted. The patient had continued proximal interphalangeal joint contracture with the development of pain and bowstringing over the next 7 months, and MRI (magnetic resonance imaging) demonstrated A1, A2, C1, and A3 pulley disruption, which required pulley reconstruction. The tendon showed signs of chronic inflammation. It has been noted that fluorinated corticosteroids, which include triamcinolone, are associated with tendon rupture5. It is thought that injections into soft tissues cause atrophy of collagenous tissue5. Betamethasone is also fluorinated, while methylprednisolone is non-fluorinated.
Plantar fascia rupture is another concern after injections into the heel. Kim et al. reported this complication in 3 of 120 patients after an average of 2.67 injections38. While conservative treatment can lead to symptomatic improvement, long-lasting complications and symptoms after rupture, such as metatarsal stress fractures, arch pain, and foot weakness, have been reported39.
Corticosteroids induce both skeletal muscle and hepatic insulin resistance, leading to hyperglycemia40. Additionally, glucocorticoids suppress the hypothalamic-pituitary-adrenal axis, thereby impairing glycemic control and decreasing serum cortisol levels40. Most reports have described an increase in blood glucose at 1 day after injection41-45. However, some investigators have found an immediate response, with increased blood glucose as early as 1 hour after injection46,47. While the earliest return to baseline occurred between 18 and 72 hours in 1 study46, other investigators have found a prolonged response lasting between 5 and 7 days41-43,47,48. Studies have not demonstrated any clear association between the amount, location, or formulation of corticosteroid injections and the resultant increase in blood glucose48. This is certainly an area of future research to be explored.
This elevation is particularly important in patients with diabetes43 and can be striking in those with insulin-dependent diabetes. Stepan et al., in a study of 40 patients with diabetes who received a methylprednisolone injection in the hand or wrist for the treatment of de Quervain tenosynovitis, trigger finger, osteoarthritis, or carpal tunnel syndrome, reported that patients with type-1 diabetes or requiring insulin were more likely to have a >50 mg/dL increase in blood glucose on the first day after injection45. Wang and Hutchinson, in a study of 18 diabetic patients, reported an average increase of 73% in blood glucose levels on the first morning after corticosteroid injection42. Conversely, while Catalano et al. also discovered a significant increase (p = 0.046), they concluded that the increase was not clinically important, averaging 14.2 and 9.7 mg/dL on days 1 and 5 after the injection48. However, patients with a pancreatic islet autotransplant can experience a clinical impact on glycemic control after injection. Ngo et al. reported on 4 patients who had total pancreatectomy and intraportal islet autotransplantation for the treatment of severe, chronic pancreatitis40. One patient, a 56-year-old woman, required 3 corticosteroid injections for the treatment of shoulder and neck pain within a 3-month period. The HbA1c (hemoglobin A1c) level increased such that she required the addition of an oral antidiabetic medication for 2 months before resolution.
Local soft-tissue reactions, such as subcutaneous fat atrophy and skin hypopigmentation, have been reported in 3% to 26% of patients after corticosteroid injection7,20,29,31,34,39. Of note, fat atrophy can occur in the plantar fat pad after injections in the foot for the treatment of plantar fasciitis. Injections for the treatment of Morton neuroma that cause fat pad atrophy could, therefore, seemingly increase, rather than decrease, foot pain49. Interestingly, Pavone et al. reported a skin dyschromia reaction in 3 of 23 children who received serial methylprednisolone acetate injections into the proximal part of the humerus for the treatment of unicameral bone cysts50. While these reactions tend to resolve within 3 to 6 months30,31,50,51, Kumar and Newman reported permanent reactions in 2 patients who had received an injection for the treatment of lateral epicondylitis30. Furthermore, Loopik et al. reported the cases of 2 patients with both fat atrophy and skin depigmentation after injection for the treatment of medial tibial stress syndrome; both patients required plastic surgery consultation for lipofiling of the lesions52. Similar to tendon ruptures, skin atrophy is thought to occur more frequently with use of fluorinated corticosteroids5.
Central Serous Chorioretinopathy
A lesser-known side effect of corticosteroid injections is central serous chorioretinopathy (CSCR). CSCR is an idiopathic detachment of the macula, which may be due to a corticosteroid-induced effect on the permeability of the choriocapillaries53 leading to the disruption of choroidal circulation54. Patients may present with decreased or distorted visual acuity and abnormal color vision55. Carvalho-Recchia et al., in a study of 50 patients with central serous chorioretinopathy, reported that 26 patients had a history of corticosteroid use and that 3 of those 26 patients had had a previous intra-articular injection; this rate of corticosteroid use was significantly higher than that in the control group (p < 0.0001)54. Notably, case reports have shown that CSCR can occur after injection in a variety of locations, including the knee56, shoulder53,57, heel58, and wrist59. Symptoms have been reported to appear as early as 36 hours after injection56 and as long as 5 to 7 days after injection53,57,58. Ophthalmological examination demonstrating serous retinal detachment of the macula also can confirm the diagnosis55,57-59. There is no medical treatment that has proven to be effective for CSCR; however, acetazolamide has been suggested to hasten the resolution of subretinal fluid, and direct focal laser photocoagulation, with low-intensity laser burns to the leakage site, has been shown to abbreviate the disease course but to have no effect on final visual acuity or recurrence rate55. Given the good prognosis with spontaneous recovery, treatment is not always indicated. Indeed, improvement occurred in all cases that we reviewed, usually starting within 2 to 3 weeks after injection57, with full resolution by 3 months53,56,58. Vision also may continue to improve up to a year55.
Anupama et al., in a case report on a patient who developed CSCR after corticosteroid treatment, noted that the patient had no history of eye problems58. Han et al.60, in a prospective pilot study, examined the underlying patient characteristics that may predispose to CSCR. Patients with no previous eye problems who required the use of high-dose corticosteroids were examined before and after corticosteroid treatments. Only 1 patient developed signs of CSCR on ophthalmological examination. That patient showed a trend toward having a higher baseline choroidal thickness and an increase in choroidal thickness after treatment.
A very rare, although worrisome, complication after injection of corticosteroid is osteonecrosis. Kontovazenitis et al.61 reported on an 80-year-old woman with radiographic evidence of chondrocalcinosis and arthritis who was managed with 3 intra-articular injections within 8 months. The patient had continued pain, effusion, and limping, and repeat radiographs revealed osteonecrosis of the medial femoral condyle, which was validated intraoperatively during total knee arthroplasty. Miyanishi et al., in a rabbit model of osteonecrosis after an intramuscular corticosteroid injection, reported increases in intraosseous pressure and the size of bone marrow fat cells and decreases in the blood flow rate and the number of bone marrow hematopoietic cells62. The authors suggested that the enlargement of marrow fat cells causes ischemia, either from compression related to the increased pressure or from intravascular fat emboli, resulting in osteonecrosis. In fact, treatment with lovastatin63,64 (which inhibits adipogenesis) and enoxaparin64 reduced the prevalence of osteonecrosis in rabbit models. While Kerachian et al.65 also noted increased hypercoagulability and decreased blood flow, they reported that these findings are the results of multiple pathways, such as the ability of corticosteroids to induce platelet thrombin formation and to regulate vasoactive substances. Additionally, they reported that corticosteroids increased apoptosis of osteocytes and osteoblasts and decreased angiogenesis. The role of corticosteroids in causing osteonecrosis needs to be further studied, as does the toxicity of systemic corticosteroids versus intra-articular injections.
There is also concern that intra-articular injections, aimed at reducing pain resulting from arthritis, may cause further destruction of the joint. In vitro studies of monolayer chondrocyte cultures have demonstrated various formulations of corticosteroids to be chondrotoxic66-69 in a dose-dependent and time-dependent manner70-72. It appears that chondrocyte death initially results from cell necrosis and, over the next several days, is more a result of apoptosis70. Comparison studies have indicated that betamethasone appears to be the most chondrotoxic corticosteroid68,69,71.
Ex vivo studies involving the use of cartilage plugs have shown that only the superficial layer is affected, acting as protection to deeper layers and allowing for increased viability of chondrocytes66,67. Prednisolone affected the superficial third of hyaline cartilage, and betamethasone had only minor cytotoxic activity66.
Verifying these results in vivo is important in order to fully understand the joint reaction to corticosteroids. Huppertz et al., in a study of 21 children (median age, 10 years) who were evaluated with MRI before and 13 months after injection with triamcinolone, reported no evidence of cartilage injury73. Raynauld et al., in a study of 68 patients who received either triamcinolone or saline solution knee injections every 3 months over 2 years, reported no radiographic difference in mean joint space width, suggesting that, in clinical practice, corticosteroids do not lead to progression of joint destruction13.
It is possible that this toxicity is not related to the corticosteroids themselves but, in fact, to the additives. Davis et al.74 found that betamethasone acetate and betamethasone sodium phosphate were not cytotoxic on their own. Conversely, benzalkonium chloride, a preservative agent used against bacterial contamination, caused cell death in a dose-dependent manner, with results occurring well below the concentration added to betamethasone. They reported that it causes disruption of intermolecular interactions and the cellular membrane bilayers, leading to leakage of cellular contents and cell death. An agent of the same class is added to methylprednisolone, highlighting the importance of further testing of the individual components in commercially available corticosteroids.
Chondrotoxicity from Local Anesthetic
The addition of a local anesthetic can provide immediate relief and diagnostic information to the physician. However, there is concern that common anesthetics such as lidocaine, bupivacaine, and ropivacaine can damage the cartilage, causing chondrolysis. This concern has become of particular interest with the increasing use of pain pumps, usually employing bupivacaine, after shoulder arthroscopy. Matsen and Papadonikolakis, in a systematic review of 213 cases of glenohumeral chondrolysis that had been reported since 2004, found that 79% of the cases were associated with postoperative pain pumps75. However, the joint exposure to local anesthetics is different with use of infrequent injections, and Ravnihar et al., in a study of cartilage biopsy specimens from patients who had been managed with an intra-articular injection of 2% lidocaine, spinal anesthesia, or general anesthesia at the time of knee arthroscopy, reported no differences between the groups in terms of cartilage morphology76. Still, a study examining intra-articular injections of bupivacaine into the knees of rats demonstrated a 50% reduction of chondrocytes at 6 months77, and multiple in vitro studies have demonstrated that there appears to be a dose-dependent and time-dependent effect of local anesthetic on chondrotoxicity78-80. It has been suggested that local anesthetics influence potassium, calcium, and sodium channels to cause mitochondrial dysfunction and subsequent cell necrosis or apoptosis81. Grishko et al. reported that they were able to overcome the potassium blockage effects and resultant mitochondrial dysfunction of lidocaine by bathing samples in high potassium solutions78.
Ropivacaine appears to be the least chondrotoxic local anesthetic, sometimes not causing any detriment compared with controls66,78,81-84. Additionally, some investigators have found bupivacaine to be less chondrotoxic than lidocaine78,84. This decreased cytotoxicity may be because bupivacaine and ropivacaine are highly lipophilic molecules, whereas lidocaine is only slightly lipophilic83. The addition of epinephrine has caused conflicting results. Some investigators have found a harmful effect on chondrocyte viability84 and chondrocyte membrane integrity79, but Jacobs et al.80 found that epinephrine was protective against cell death. Syed et al. reported that the addition of a buffer to maintain the physiologic pH of synovial fluid was not beneficial67.
Similar to the results associated with corticosteroids, studies have shown that the superficial layer is protective and improves the overall viability of chondrocytes66,67. Matsen and Papadonikolakis noted that shoulder arthroscopic procedures involving the use of ≥1 suture anchors and postoperative pain pumps have consistently been reported to increase chondrolysis75. The authors suggested that this breach of the superficial layer facilitates diffusion of the anesthetic into the substance of the cartilage75.
Combining corticosteroids and a local anesthetic may have a synergistic effect on chondrocyte death67,69,70. In several studies, the combination of betamethasone with either lidocaine66,69 or bupivacaine69 has been found to cause the most apoptosis in chondrocyte cultures. Farkas et al. noted that the combination of betamethasone and lidocaine affected only the superficial aspect of ex vivo osteochondral plugs but that the combination of prednisolone and lidocaine caused chondrocyte death across the whole depth of the cartilage sample66. Interestingly, while synergistic effects occurred when methylprednisolone or triamcinolone were mixed with lidocaine, this was not the case when these same agents were mixed with bupivacaine69. Furthermore, using less-concentrated doses of the corticosteroid and local anesthetic has been shown in rat models to reduce chondrotoxicity to nonsignificant levels (p = 0.738)72. This finding suggests that a balance may exist between mitigating symptoms in patients and reducing any potential harm.
Corticosteroid injections are frequently used not only to diminish pain but also with the aim of curing certain conditions. With their widespread use, mostly done in an office or other outpatient setting, it can be easy to dismiss any potential complications. While most complications are rare and minor, it is still important to discuss these potential complications with patients prior to injection. Effects on blood glucose are of particular concern to patients with diabetes, who may feel the immediate effects of injections, both at the site of pain and with increasing blood glucose. Even relatively minor complications, such as fat atrophy and skin hypopigmentation, can require additional medical consultations and surgery. However, it is difficult to strike a balance of notifying patients of the risks without overwhelming them with concern, especially as the purpose of corticosteroid injections is to provide relief. As practitioners, it is therefore most prudent to administer corticosteroid injections in a way that most likely avoids complications. As more clinical guidelines are released, it is likely that the role of corticosteroids will be addressed, particularly when one considers their common use as a treatment modality for many orthopaedic conditions. Future guidelines will provide a useful adjunct in helping physicians to judge the best use and indications for corticosteroid injections in their practices. However, when the research is inconclusive, more consideration must be given to the experience of the physician and discussions with the patient regarding the patient’s goals, his or her health status, and alternative options and associated complications.
Investigation performed at the Rutgers-Robert Wood Johnson University Hospital, New Brunswick, New Jersey
Disclosure: No external funding was provided for this article. The Disclosure of Potential Conflicts of Interest forms are provided with the online version of the article.
- Copyright © 2016 by The Journal of Bone and Joint Surgery, Incorporated