Injection of the hip for osteoarthritis is typically performed for diagnostic or therapeutic reasons. For diagnostic purposes, an intraarticular anesthetic is used to distinguish intrinsic sources of hip pain from extrinsic sources of pain such as spinal pathology,5,6,12 or to differentiate knee pain potentially referred by hip disease.11 Therapeutic injections for arthritis of the hip contain steroids or viscosupplements.3,7,14,16,20 These therapeutic injections are used less frequently for the treatment of arthritis of the hip when compared to their common use for the treatment of arthritis of the knee.21 This difference is partly because of the increased technical difficulty of gaining access to the hip without fluoroscopic or ultrasound guidance,3,10,16 and the unpredictable duration of beneficial pain relief.18 As a consequence, steroid injections to the hip are not currently recommended in the treatment guidelines provided by the American College of Rheumatology1 for osteoarthritis (OA) of the hip.
However, we have noted increased rates of therapeutic intraarticular steroid injection before our patients are referred for total hip arthroplasty (THA). Any form of systemic steroid therapy taken for greater than 1 week within 1 year before THA increases the risk of prosthetic joint infection two-fold compared with normal control subjects.2 In a recent report, of 40 patients who had THA after an intraarticular steroid injection, 10% had a deep peri- prosthetic joint infection develop.11 This risk was different when compared with a matched group of patients who did not have a steroid injection to the hip. Because of the small study size, the authors were unable to determine whether there was an additive effect with multiple injections, or the effect of the time lapse between the injection and the THA. This information heightened negative suspicions about the association of THA after intraarticular steroid injection and development of superficial or deep periprosthetic infections.
We hypothesized that the administration of an intraarticular steroid injection into an osteoarthritic hip within 1 year of subsequent primary THA would increase the rate of superficial and deep periprosthetic infection when compared with a matched control group who did not receive a preoperative intraarticular steroid injection.
MATERIALS AND METHODS
We created retrospective matched cohorts from the medical records for 3848 THAs for patients with OA of the hip between January 1998 and May 2002. Cases were identified using a computerized database compiled by the Mayo Clinic Total Joint Registry, and the Mayo Clinic Master Diagnostic Index9,15 after obtaining approval from our institutional review board. The rate of therapeutic intraarticular steroid injections one year prior to THA was 5.8% of the 3848 primary THAs. This same group of patients, (Group 1: 217 patients, 224 THA) accounted for 35.7% (224/628) of the therapeutic intraarticular steroid injections performed to treat OA of the hip during the same period. Of these procedures, 224 (5.8%) hips were implanted in 217 patients who received an ipsilateral hip intraarticular steroid injection preoperatively. The 224 THAs in 217 patients (Group 1) were matched with a control group (Group 2) of 224 primary THAs in 220 patients who did not receive intraarticular steroid injections. The cohorts were matched according to gender, age, and surgeon. The joint registry does not contain information on systemic diseases or immunosuppressive medications. Consequently, this information could not be incorporated into the matching process. Group 1 included 93 men and 131 women with a mean age of 70 years (range, 35-94 years) at the time of THA. Group 2 included 92 men and 132 women with a mean age of 69 years (range, 41-92 years) at the time of THA. The preoperative diagnosis for all patients was OA (Table 1). Exclusion criteria for both groups included a diagnosis of inflammatory arthritis, connective tissue disorders, any history of acetabular or femoral fracture, any previous surgery on the index hip, or tumor of the acetabulum or femur.
The mean time between steroid injection and THA was 112 days (standard deviation [SD] = 81 days). All injections were performed by members of our radiology department within 1 year before the THA using standard aseptic protocols. Hip penetration was confirmed fluoroscopically by the installation of radiopaque dye before steroid administration. The type and amount of intraarticular steroid dispensed was left to the discretion of the radiologist performing the injection. The average dose of intraarticular steroid dispensed was 20.47 mg (6-40 mg).
The main outcome measure was implant survival time. All patients were followed at 3 months, 1 year, 2 years, and 5 years in accordance with guidelines established by the joint registry. The patients were observed until THA revision for instability, loosening, implant resection for deep infection, or death. The minimum potential followup was 2 years. The average duration of clinical followup was 2.7 years (SD = 1.4 years) in Group 1 and 2.6 years (SD = 1.6 years) in Group 2. Superficial infections were defined as any wound infection that did not penetrate the deep fascia, and included any patients with persistent postoperative wound drainage, superficial wound dehiscence, or suture abscess formation. There were 25 deaths unrelated to the THA (12 in Group 1 and 13 in Group 2) during the study period.
Descriptive statistics were reported as a number (percent) for discrete values and as a mean (SD) for continuous values. The cumulative risk of deep and superficial infections were estimated separately for hips replaced using an intraarticular steroid injection and for hips replaced without this injection by using the Kaplan-Meier survival method. The Cox proportional hazard survival method adjusting for correlated data (ie, two hips in patients with bilateral THAs) was used to assess the difference in survivals between the two groups. This hazard model takes into account the timing of injection and the length of followup. The hazard ratios from these models were reported in addition to 95% confidence intervals. For example, a hazard ratio of 1 meant the risk of deep infection was equivalent for the two groups. A hazard ratio of 2 meant the risk in Group 1 was twice that of Group 2, and a hazard ratio of 0.8 would mean that the risk in Group 1 is 20% less than the risk in Group 2. Significance was set at a p value of 0.05.
Receiving an ipsilateral intraarticular steroid injection within 1 year before THA did not influence the postoperative rates of deep or superficial infection. There were 11 superficial infections in Group 1 (4.9%), and eight superficial infections in Group 2 (3.6%). This represents a hazard ratio of 1.5 (95% CI, 0.6, 3.6). The cumulative risk of superficial infection at 5 years was 4.5% (95% CI, 2.4, 8.4) in Group 1 and 3.7% (95% CI, 1.1, 6.1) Group 2.
Nineteen patients in Groups 1 and 2 had superficial infections develop. Fifteen of these patients were treated with local wound care and seven of these 15 patients were also treated with a course of oral antibiotics. The remaining four patients underwent reoperations. Two patients had surgical removal of retained drains, and two patients had their wounds incised followed by irrigation and débridement with documentation of fascial integrity. None of the 19 patients diagnosed with superficial wound infections had a deep wound infection develop at final followup.
Three patients in Group 1(3/224) and one patient in Group 2 (1/224) had deep infection of the hip develop at a mean of 1.69 years (0.25-3.85 years) and 1.51 years, respectively. This represents a hazard ratio of 3 (95% CI, 0.3, 29.8). Of these four infections, two were methicillin- resistant Staphylococcus aureus, one was Escherichia coli, and one had no growth from cultures obtained at the time of débridement. Three infections were treated with two-stage exchange and one by débridement with prosthesis retention and chronic suppression with oral antibiotics. Half of the patients who had deep infection develop also had a chronic medical condition. One patient had a tracheostomy for severe obstructive sleep apnea, and was on chronic anticoagulation with Coumadin (Bristol-Myers Squibb, Plainsboro, NJ) for atrial fibrillation and deep venous thrombosis and/or pulmonary embolism. The other patient had Type II diabetes mellitus. The cumulative risk of deep infection at 5 years was 3% (95% CI, 0%, 7%) in Group 1 and 0.6% (95% CI, 0%, 1.7%) in Group 2. Two of the four deep infections were preceded by multiple recurrent dislocations; both occurred in Group 1.
We found no association between the average time from intraarticular steroid injection to primary THA and the development of superficial or deep infection. The mean time between steroid injection and THA was 112 days (SD = 81 days) for all Group 1 patients and in those patients who had a superficial infection develop the mean time also was 112 days (SD = 94 days). However, the mean time from injection to THA was 44 days (SD = 23 days) for those patients who had a deep infection develop (a probability value could not be calculated because of the limited numbers of patients with deep infections).
The revision rate for noninfected cases (instability and aseptic loosening) was similar in both groups: 5/224 for Group 1 and 4/224 or Group 2. There were six dislocations in the Group 1 and five in the control group. The mean time to dislocation was 1.41 years (range, 0.02-3.13 years) and 1.54 years (range, 0.02-5), respectively. The time from steroid injection to THA in this subset of patients was 79 days (SD = 69; p = 0.35). Four patients demonstrated hip stability with conservative treatment. The other seven patients had multiple dislocations and underwent reoperation to improve stability. Of the seven patients who underwent reoperation, four were in Group I and three were in Group 2. Both groups had one patient (1/224) undergo revision arthroplasty for aseptic loosening of the femoral and acetabular components at 4.17 years in Group 1 and 1.58 years in Group 2. The time from steroid injection to primary THA was 97 days.
The question raised in this investigation was whether the administration of an intraarticular steroid injection into an osteoarthritic hip within 1 year of subsequent primary THA would demonstrate an increased rate of superficial and deep periprosthetic infection when compared with a matched control group who did not receive a preoperative intraarticular steroid injection.
This investigation has several limitations. First, the small sample size in this study imparts low statistical power. Second, the retrospective study design involved multiple surgeons in the treatment of these patients. Third, multiple radiologists at our institution performed the intraarticular steroid injection, and the type and amount of steroid compound injected into the hip were not controlled by a specific protocol. Lastly, medical conditions and immunosuppressive medications were not incorporated into the matching process. Thus, the data must be interpreted as preliminary but suggesting the need for further study.
The diagnostic value of injecting 0.5% Marcaine (AstraZeneca Pharmaceuticals LP, Wilmington, DE) to the hip under fluoroscopic guidance in patients with considerable pain but minimal radiographic changes and/or concomitant spinal pathology has been demonstrated.5 Diagnostic intraarticular Marcaine injections have become accepted in the orthopaedic community as a reliable test to aid in the clarification of the source of pain when the etiology is indeterminate.
The safety and efficacy of intraarticular steroid injection to treat OA of the hip is not as clearly defined. Several reports have documented the occurrence of septic arthritis following an intraarticular steroid injection in the hip joint.4,13,17,19 Five hundred ten patients with OA of the hip who were injected with 3 mL triamcinolone acetonide (40 mg/mL) using fluoroscopic guidance did not have complications develop.16 Of the two specific end-points addressed, 198 patients (38.8%) had pain relief that lasted at least 8 weeks, and 154 patients (30.1%) had an increase in internal rotation of at least 10°.16 Also, patients with moderate and mild OA of the hip had a better response to the injection than patients with severe OA.16
In another study of 11 patients injected with 1 mL of 40 mg/mL triamcinolone under ultrasound guidance, there were also no complications.10 Using a 100 mm visual analog scale (VAS), only 55% of patients had more than a 15 mm score reduction at 2 weeks, which decreased to 33% and 38% at 6 weeks and 12 weeks, respectively. None of the patients demonstrated improved range of motion (ROM).10 Although these studies demonstrate steroids can be safely injected to the hip using fluoroscopic and ultrasound guidance, it is also clear that the duration of beneficial response is relatively brief.10,16
Systemic steroid therapy increases the risk of prosthetic joint infection two-fold compared with normal control subjects.2 Patients receiving any form of systemic steroid therapy for greater than 1 week within 1 year before THA have a two-fold increased risk of prosthetic joint infection compared with normal control subjects.2
In a recent study, a group of 40 patients who had intraarticular steroid injections and a subsequent THA were compared with 40 control patients who had THA without receiving steroid injections.11 These patients included several with inflammatory arthritis. There were five revisions in the injection group (four for deep infection) compared with no revisions in the control group.11 Six other patients who received injections underwent investigation for infection because of persistent problems in the hip compared with only one patient in the control group.11 These deep infections contributed to an overall revision rate of 12.5% (five of 40 THAs) in this group compared with a 1.02% revision rate (10 of 979 THAs) of concomitant primary THAs performed at their hospital over the same time period.11 The mean number of injections was 1.43 (SD = 0.25) occurring at a mean of 11 months (range, 0.5-43 months) before the THA.11 There was no difference in the timing of the injection before THA in patients who had deep wound infections and those who did not.11
Our study design was limited to patients with a diagnosis of primary OA and patients who received an injection within 1 year before their THA. It was developed to create a uniform group of patients to potentially assess the effect of therapeutic intraarticular steroid injection before THA. The rate of therapeutic intraarticular steroid injection within 1 year of subsequent THA was 5.8% (224/3848) in this study. In contrast with the previous study,11 our results suggest receiving an ipsilateral intraarticular steroid injection within 1 year before THA did not affect postoperative rates of deep or superficial infection. Historically, at our institution between 1969 and 1996, deep infection occurred in 1.3% of 23,519 primary total hip arthroplasties.8 This is equivalent to the rate of deep infection demonstrated in this study (1.3%; 3/224). There was no association between the average time from injection to THA and the development of superficial or deep infection. However, when critically evaluating the data, the average time from injection to THA was less in those patients that developed a deep infection (44 days) compared to Group 1 as whole (112 days), and the hazard ratio for deep infection was 3 when comparing the patients who received an injection (Group 1) with the control patients (Group 2). This raises the concern of increased risk of deep infection when receiving a steroid injection within 6 weeks before THA.
In half of the patients who have deep infection develop, the diagnosis was preceded by multiple recurrent instability episodes. Based on the small numbers, no significance can be attributed to this observation. This suggests that recurrent instability and deep infection may be linked. The cause and effect are not known, and future prospective investigation is warranted.
We recommend using caution and judgment before performing an intraarticular steroid injection to conservatively treat primary OA of the hip if THA is being contemplated in the near future. It was not possible to determine an exact time frame based on our available data, but we suggest avoiding the use of intraarticular steroids injected in the hip for at least 2 months before THA.
The authors acknowledge Cathy Schleck and Scott Harmsen for their dedication and support.
1. American College of Rheumatology Subcommittee on Osteoarthritis. Guidelines Recommendations for the medical management of osteoarthritis of the hip and knee: 2000 update. Arthritis Rheum
2. Berbari EF, Hanssen AD, Duffy MC, Steckelberg JM, Ilstrup DM, Harmsen WS, Osmon DR. Risk factors for prosthetic joint infection: case-control study. Clin Infect Dis
3. Caglar-Yagci H, Unsal S, Yagci I, Dulgeroglu D, Ozel S. Safety and efficacy of ultrasound-guided intra-articular hylan G-F 20 injection in osteoarthritis of the hip: a pilot study. Rheumatol Int
. 2005;25: 341-344.
4. Chazerain P, Rolland D, Cordonnier C, Ziza JM. Septic hip arthritis after multiple injections into the joint of hyaluronate and glucocorticoid. Rev Rheum Engl Ed
5. Crawford RW, Gie GA, Ling RS, Murray DW. Diagnostic value of intra-articular anaesthetic in primary osteoarthritis of the hip. J Bone Joint Surg Br
6. Faraj AA, Kumaraguru P, Kosygan K. Intra-articular bupivacaine hip injection in differentiation of coxarthrosis from referred thigh pain: a 10 year study. Acta Orthop Belg
7. Flanagan J, Casale FF, Thomas TL, Desai KB. Intra-articular injection for pain relief in patients awaiting hip replacement. Ann R Coll Surg Engl
8. Hanssen AD, Rand JA. Evaluation and treatment of infection at the site of a total hip or knee arthroplasty. J Bone Joint Surg Am
9. Ilstrup DM. Statistical considerations and the science of surveil- lance. In Reconstructive Surgery of the Joints. New York: Churchill-Livingstone; 1996:225-301.
10. Karim Z, Brown AK, Quinn M, Wakefield RJ, Conaghan PG, Emery P, O'Connor PJ. Ultrasound-guided steroid injections in the treatment of hip osteoarthritis: comment on the letter by Margules. Arthritis Rheum
. 2004;50:338-339; author reply 339-340.
11. Kaspar S. de V de Beer J. Infection in hip arthroplasty after previous injection of steroid. J Bone Joint Surg Br
12. Kleiner JB, Thorne RP, Curd JG. The value of bupivacaine hip injection in the differentiation of coxarthrosis from lower extremity neuropathy. J Rheumatol
13. Kortelainen ML, Sarkioja T. Fatal complications of intramuscular and intra-articular injections. Z Rechtsmed
14. Kullenberg B, Runesson R, Tuvhag R, Olsson C, Resch S. Intraarticular corticosteroid injection: pain relief in osteoarthritis of the hip? J Rheumatol
15. Kurland LT, Molgaard CA. The patient record in epidemiology. Sci Am
16. Margules KR. Fluoroscopically directed steroid instillation in the treatment of hip osteoarthritis: safety and efficacy in 510 cases. Arthritis Rheum
. 2001;44:2449-2450; author reply 2455-2446.
17. Nallamshetty L, Buchowski JM, Nazarian LA, Narula S, Musto M, Ahn NU, Frassica FJ. Septic arthritis of the hip following cortisone injection: case report and review of the literature. Clin Imaging
18. Neustadt DH. Local steroid injections: comment on the American College of Rheumatology guidelines for the management of osteo- arthritis of the hip and on the letter by Swezey. Arthritis Rheum
19. Pal B, Morris J. Perceived risks of joint infection following intraarticular corticosteroid injections: a survey of rheumatologists. Clin Rheumatol
20. Plant MJ, Borg AA, Dziedzic K, Saklatvala J, Dawes PT. Radio- graphic patterns and response to corticosteroid hip injection. Ann Rheum Dis
21. Roddy E, Doherty M. Guidelines for management of osteoarthritis published by the American College of Rheumatology and the European League Against Rheumatism: why are they so different? Rheum Dis Clin North Am