Accuracy of Diagnostic Injection in Differentiating Source of Atypical Hip Pain

From thYork, NewYork, New

SubmitNo benReprint

New York© 20100883-5doi:10.1

The Journal of Arthroplasty Vol. 25 No. 6 Suppl. 1 2010

Accuracy of Diagnostic Injection inDifferentiating Source of Atypical Hip Pain

Ajit J. Deshmukh, MD,* Raman R. Thakur, MD,* Amrit Goyal, MD,*Devon A. Klein, MD,y Amar S. Ranawat, MD,* and Jose A. Rodriguez, MD*

Abstract: It is not uncommon to encounter patients with atypical hip or lower extremity pain, ill-defined clinicoradiological features, and concomitant hip and lumbar spine arthritis. The purpose ofthis study is to present our experience using the response resulting from a combined anesthetic-steroid hip injection for treatment selection in these patients. A retrospective analysis of 204consecutive diagnostic hip injections was undertaken. Patient charts were scrutinized for outcomesof injection and treatment. Our findings suggest that the relief of symptoms following injection oflocal anesthetic and steroid into the hip joint has a sensitivity of 91.5%, specificity and positivepredictive value of 100%, and negative predictive value of 84.6% for response to total hiparthroplasty. We thereby believe that this is a reliable test with low morbidity and can predict thepotential benefit of total hip arthroplasty in this diagnostically challenging group of patients.Keywords: diagnostic, anesthetic, steroid, hip injection.© 2010 Elsevier Inc. All rights reserved.

The “hip region” constitutes the groin, buttock, upperlateral thigh, greater trochanteric area, and the iliac crest.Pain originating from various sources may be perceivedhere and includes the hip joint, lumbosacral spine,sacroiliac joints, pubis, and soft tissue sources such astrochanteric bursitis, hip abductor dysfunction, andinguinal hernia. The prevalence of hip osteoarthritis isknown to increase with age, affecting approximately 4%of the population older than 65 years [1]. Ten percent to15% of patients may exhibit simultaneous involvementof the lumbar spine and hip(s) [2]. Careful history taking,physical examination, and plain radiographs are believedto provide crucial information in the assessment ofindividuals with hip disease. However, a diagnosticdilemma can arise in patients with atypical symptomsand signs. Even the presence of radiographic hip or spinearthritis does not always correlate with the presence ofsymptoms [3,4].It is a challenging clinical situation when history, clinical

examination, and plain radiography fail to locate the exactorigin of hip pain. The quantification of symptoms inconcomitant hip and spine disease is also vital andparticularly relevant if the management includes a major

e *Department of Orthopaedic Surgery, Lenox Hill Hospital, NewYork; and yDepartment of Radiology, Lenox Hill Hospital, NewYork.ted July 14, 2009; accepted April 1, 2010.efits or funds were received in support of the study.requests: Ajit J. Deshmukh, MD, 130 E, 77th St, 11th Floor,, NY 10075.Elsevier Inc. All rights reserved.

403/2506-0026$36.00/0016/j.arth.2010.04.015

129

reconstructive surgery such as hip arthroplasty. Additionaldiagnostic workup includes an anesthetic hip injection thathelps differentiate the source of pain [4-8]. Most literaturereports use of local anesthetic injections [4-7]. However, ithas been our practice to use a combination of localanesthetic and steroid. The aim of this study was to assessthe accuracy of anesthetic-steroid hip injection whenapplied as a diagnostic tool for hip arthritis in this clinicallychallenging group of patients.

Patients and MethodsAfter obtaining Institutional Review Board approval,

we reviewed the clinical and radiographic records of allpatients under care of the Arthroplasty service whounderwent a hip injection between October 2005 andOctober 2008 at our institution. Of the total 267 patients,204 individuals were included into the study becausethey satisfied at least one of the following conditions.

1. Pain in the hip region of uncertain etiology with/without radiating knee pain of minimum 6 months'duration

2. Subtle degenerative changes in the hip joint onplain radiographs

3. Concurrent hip and lumbar spine arthritis and4. Absence of localizing physical signs with clinical

examination such as Stinchfield test (resisted hipflexion test).

Patients were excluded if they had isolated labral tearsof the hip or were undergoing therapeutic hip injectionsuch as those awaiting hip arthroplasty.

130 The Journal of Arthroplasty Vol. 25 No. 6 Suppl. 1 September 2010

All patients were first assessed with detailed history,clinical examination, and anteroposterior and lateralradiographs of the hips and lumbar spine. If the diagnosiswas in doubt or if relative quantification of pain wasdifficult, patients were offered a diagnostic hip injection.

Hip Arthrogram TechniqueAll hip injections were performed in a dedicated

radiology suite under strict sterile conditions by thesame musculoskeletal radiologist with several years ofexperience and more than 2000 injections. Skinpreparation was undertaken using povidone iodinesolution, and the area was draped. The skin wasinfiltrated with 1% lidocaine. Under fluoroscopicguidance, a 20-gauge spinal needle was advanced intothe hip joint from anterolateral side targeting theanterior surface of lateral femoral neck. Once the lossof resistance was felt, intraarticular position of theneedle was confirmed by injecting 3 mL Omnipaque240 (GE Healthcare Inc, Princeton, NJ); and a spotimage was taken to document location (Fig. 1). In onepatient, because of previous allergy to radioopaque dye,room air was injected to confirm needle placement. Amixture of 5 mL 0.5% bupivacaine/Sensorcaine (Astra-Zeneca, Wilmington, DE) and 1 mL (80 mg) ofmethylprednisolone (either Depo-Medrol [Pfizer Inc,New York, NY] or generic) was then injected into thehip joint.

Analysis of Response to InjectionAfter the injection, patients were observed for 30

minutes; and feedback about pain relief was documentedin the radiology report. The response to injection was

Fig. 1. Spot radiograph demonstrating intraarticular pooling ofthe dye, confirming needle placement into the hip joint.

analyzed in terms of percentage relief of pain, with a“positive response” meaning more than 50% pain relieffrom the preinjection pain level and “negative response”meaning less than 50% pain relief. Patients wereencouraged to ambulate and carry out all routineactivities. All patients were followed up by the Arthro-plasty service 2 weeks after injection. Negative respon-ders were interviewed about pain relief subsequently (ie,after the 30-minute observation period in the radiologysuite) and categorized as “delayed positives” if they hadresponded positively within 2 weeks. These patients wereadded to the “positive response” group. The “negativeresponse” group was investigated further to diagnose thesource of pain, and some were referred to the spineservice. The outcome of total hip arthroplasty (THA) interms of pain relief was determined at a minimum 6months of follow-up. Harris Hip Score was used todocument preoperative disability and improvement aftersurgery [9].A standard 2 × 2 table was used to calculate sensitivity,

specificity, and positive and negative predictive value ofthe test (Table 1).

ResultsThere were no complications of injection in any of

the patients. Of 204 individuals, there were 128women and 76 men, with a mean age of 65.40years (range, 31-84 years). One hundred fifty-twopatients (74.5%) had a positive response (127immediate and 25 delayed), and 52 (25.5%) had anegative response to injection. In the positive re-sponse group (n = 152), there were 97 women and 55men, with a mean age of 65.81 years. Eighty-six(56.6%) of 152 positive responders underwent pri-mary uncemented THA, and all had good pain reliefat minimum 6 months of follow-up (true positives[TPs]). Final diagnosis at surgery was osteoarthritis inall cases. The Harris Hip Score improved in this groupfrom a mean of 56.69 preoperatively to 88.86postoperatively at 6 months. Out of the remaining66 patients in the positive response group; 28 wereawaiting surgery, 21 deferred surgery and requestedrepeat intraarticular injections that were provided tothem, 12 were being treated conservatively, and 5were lost to follow-up.The negative response group (n = 52) was composed

of 32 women and 20 men, with an average age of

Table 1. Two × Two Table

“Gold Standard” Result (Relief After THA forHip/Interventions for Other Diagnoses)

Hip Disease + Hip Disease −

Test result (hip injection)+ (TP) 86 (FP) 0− (FN) 8 (TN) 44

Table 2. Review of Literature

Study andPublication Year Injection Content and Quantity

Method of Confirmationof Injection Placement

Total No. ofCases (N) Sensitivity Specificity

PositivePredictiveValue

NegativePredictiveValue

Kleiner et al,1991

10 mL of 0.25% bupivacaine None (fluoroscopyonly in 2 cases)

18 88% 100% 100% 33%

Crawford et al,1997

10 mL of 0.5% bupivacaine Fluoroscopy 42 87%-100% 80%-100% 96%-100% 44%

Faraj et al,2003

0.5% bupivacaine (24 patients)and 0.5% bupivacaine +triamcinolone (23 patients)

Fluoroscopy 47 88% 100% 100% 73%-85%

Illgen et al,2006

3.5 mL 2% lidocaine +3.5 mL 0.5% bupivacaine

Fluoroscopy 34 95.2% 87.5% 95.2% 87.5%

Pateder andHungerford,2007

9 mL 1%lidocaine + 1 mL (40 mg)triamcinolone

Fluoroscopy 83 100% 81% 97% 100%

Differentiating Source of Atypical Hip Pain � Deshmukh et al 131

63.86 years. These were thoroughly investigated forother sources of pain. Forty-three patients (82.7%)benefited from treatment of alternative diagnosis.These included 34 patients in whom the lumbarspine was identified as the predominant source ofdiscomfort and treated successfully (true negatives[TNs]), 3 patients with trochanteric bursitis (whichwas transiently relieved with local steroid injection), 2with rheumatic diathesis, 2 with sacroiliitis, and 2 withhip abductor dysfunction.The remaining 9 patients (17.3%) of the negative

response group underwent THA because of overwhelm-ing suspicion of hip pathology (after ruling out otheretiologies), and 8 (15.4%) of these reported good relief ofsymptoms and functional improvement (false negatives[FNs]). Unfortunately, one patient (1.9%) is dissatisfiedwith surgery and continues to have the same quality andintensity of trochanteric and lateral thigh pain as beforesurgery (TN).It is interesting to note that no cases with a positive

injection response had a negative response to THA (falsepositives [FPs]).Summarizing results, we had 86 TPs, no FPs, 44

TNs, and 8 FNs. Calculations derived a sensitivity [TP/(TP + FN)] of 91.5%, a specificity [TN/(TN + FP)] anda positive predictive value [TP/(TP + FP)] of 100%each, and a negative predictive value [TN/(TN + FN)]of 84.6%.

DiscussionDegenerative joint disease is a common finding in

people older than 60 years. Hip arthritis and lumbar spinearthritis coexist in 10% to 15% of patients and oftenpresent with similar signs and symptoms [2-4,10,11]. Thecommon innervation of the hip capsule, adductorcompartment of the thigh and knee by the obturatornerve explains the presence of anteromedial thigh andknee pain in patients with hip pathology. Pain below theknee is not as easily explained and invokes the possibilityof compressive neuropathy [4].

Brown et al [3] reported signs and symptoms thathelp differentiate hip and spine disease in patientswith concomitant radiographic hip and spine arthritis,with groin, buttock, medial and lateral thigh, knee,and leg pain. Their findings suggest that although alimp, groin pain, and restricted internal rotation of thehip are more commonly associated with hip disease,they may also be present in patients with spinepathology. Birrell et al [10], in their study involving1071 patients in the age group of 45 to 84 years,found that mild to moderate osteoarthritis of the hipcan sometimes be asymptomatic, whereas Borenstein[12] reported that the same may be true for lumbarspine arthritis. Conventional radiographs appear tohave poor diagnostic ability in the detection of initialstages of hip osteoarthritis [13-15]. Santori and Villar[13] reviewed the records and radiographs of 234 hiparthroscopies and found that 32.2% of patients withnormal preoperative radiographs had evidence ofosteoarthritis at arthroscopy and that it was onlywhen both sides of the joint were involved thatosteoarthritis was evident radiographically. They notedthat painful hips in young and middle-aged patientswith normal radiographs pose a diagnostic challenge tothe orthopedic surgeon.The concept of diagnostic hip injection is not a new one.

Reports in the literature support their use in confirmingpain of hip arthritis [4-8,16,17]. Most authors used purelylocal-anesthetic injections or added steroid to some butnot all injections in their series [4-7]. To our knowledge,only the report by Pateder and Hungerford [8] hasconsistently used local anesthetic and steroid in all cases.A review of literature on diagnostic hip injection ispresented in Table 2.The results of our series indicate 100% specificity

(probability of a patient who does not have thecondition testing negative) and positive predictivevalue (probability of a patient who tests positivehaving the disease), as there were no FPs. However,the sensitivity (which is a measure of the probability of

132 The Journal of Arthroplasty Vol. 25 No. 6 Suppl. 1 September 2010

a patient with the condition testing positive) was91.5%; and the negative predictive value (probabilityof a patient who tests negative not having the disease)was 84.6%. The worst case scenario in this situation isto have a false-positive response prompting thesurgeon to perform a hip reconstructive surgery withpoor outcome leading to significant patient dissatisfac-tion. There was no such case in our study group, asthere were no FPs.The present study is different from others in the

following ways. Firstly, radioopaque dye (Omnipaque240) and fluoroscopy were used for confirmation ofintraarticular placement of the injection. Secondly,addition of methylprednisolone to local anesthetic wasdone in all cases. Thirdly, assessment of injectionresponse was done at 30 minutes and then 2 weekspostinjection. Some patients received the injection on a“good day” when they did not have much pain, andthis is probably the cause for them not respondingpositively on day 1; however, the steroid couldmaintain its effect sufficiently long for them to perceivethe extent of benefit, which they reported at the 2-week visit. Twenty-five patients reported such “delayedpositive” response and were included in the “positiveresponse” group. Total hip arthroplasty was performedon 7 of them with successful outcome in all (TPs), thusimproving sensitivity of the test. It is quite evident that,because the number of FNs was reduced, addition ofmethylprednisolone also helped improve the negativepredictive value. In addition, steroid offers simulta-neous therapeutic benefit. It was, however, surprisingto have FNs (n = 8) despite the use of anesthetic andsteroid. A literature review revealed only 90% re-sponse frequency to steroid injection in hip osteoar-thritis [18], whereas another article suggested thatpatients with atrophic pattern of osteoarthritis respondless well than those with a hypertrophic or mixedresponse [19].The merits of this study are its larger sample size as

compared with other studies and the use of fluoroscopicguidance as well as radioopaque dye to confirm correctplacement of injection. Study limitations are those of anyretrospective study and the fact that a large number ofpositive responders (n = 66) did not undergo THA andhence were excluded from statistical analysis. Thus, outof 204, only 138 subjects (86 TPs + 44 TNs + 8 FNs) couldbe ultimately included.In conclusion, our results support the role of a

diagnostic hip injection in confirming origin of painfrom an arthritic process in the hip joint. It is a valuabletool to differentiate knee pain originating from the hipfrom that originating from the knee and also todistinguish other sources of hip pain, most notably thelumbar spine. The test also predicted the potential benefitof THA in these patients. Although those with a positivetest result can be confidently expected to benefit fromTHA, the interpretation of a negative result should

however be more circumspect because of the possibilityof FNs.

AcknowledgmentWe thank Georgia Panagopoulos, PhD, for her help

with the statistics and manuscript.

References1. Lawrence JS, Bremner JM, Bier F. Osteoarthritis:

prevalence in the population and relationship betweensymptoms and x-ray changes. Ann Rheum Dis 1966;25:1.

2. Magora A. Investigation of the relation between low backpain and occupation. VII. Neurologic and orthopediccondition. Scand J Rehab Med 1975;7:146.

3. Brown MD, Gomez-Martin O, Brookfield KF, et al.Differential diagnosis of hip disease versus spine disease.Clin Orthop 2004;1:280.

4. Kleiner JB, Thorne RP, Curd JG. The value of bupivi-caine hip injection in the differentiation of coxarthrosisfrom lower extremity neuropathy. J Rheumatol 1991;18:422.

5. Crawford RW, Gie GA, Ling RSM, et al. Diagnostic value ofintra-articular anaesthetic in primary osteoarthritis of thehip. J Bone Joint Surg Br 1998;80-B:279.

6. Faraj AA, Kumaraguru P, Kosygan K. Intra-articularbupivacaine hip injection in differentiation of coxarthrosisfrom referred thigh pain: a 10 year study. Acta Orthop Belg2003;69:518.

7. Illgen RL, Honkamp NJ, Weisman MH, et al. The diagnosticand predictive value of hip anesthetic arthrograms inselected patients before total hip arthroplasty.J Arthroplasty 2006;21:724.

8. Pateder DB, Hungerford MW. Use of fluoroscopicallyguided intra-articular hip injection in differentiating thepain source in concomitant hip and lumbar spine arthritis.Am J Orthop 2007;36:591.

9. Harris WH. Traumatic arthritis of the hip after disloca-tion and acetabular fractures: treatment by moldarthroplasty. An end result study using a new methodof result evaluation. J Bone Joint Surg Am 1969;51:737.

10. Birrell F, Lunt M, Macfarlane G, et al. Association betweenpain in the hip region and radiographic changes ofosteoarthritis: results from a population based study.Rheumatology (Oxford) 2005;44:337.

11. McNamara MJ, Barrett KG, Christie MJ, et al. Lumbarspinal stenosis and lower extremity arthroplasty.J Arthroplasty 1993;8:273.

12. Borenstein D. Does osteoarthritis of the lumbar spine causechronic low back pain? Curr Pain Headache Rep 2004;8:512.

13. Santori N, Villar RN. Arthroscopic findings in the initialstages of hip osteoarthritis. Orthopedics 1999;22:405.

14. Edwards DJ, Lomas D, Villar RN. Diagnosis of the painfulhip by magnetic resonance imaging and arthroscopy.J Bone Joint Surg Br 1995;77:374.

15. McCarthy JC, Busconi B. The role of hip arthroscopy in thediagnosis and treatment of hip disease. Orthopedics 1995;18:753.

Differentiating Source of Atypical Hip Pain � Deshmukh et al 133

16. Chitre AR, Fehily MJ, Bamford DJ. Total hip replacementafter intra-articular injection of local anaesthetic andsteroid. J Bone Joint Surg Br 2007;89-B:166.

17. Kaspar J, Kaspar S, Orme C, et al. Intra-articular steroid hipinjection for osteoarthritis: a survey of orthopaedic surgeonsin Ontario. Can J Surg 2005;48:461.

18. Kullenberg B, Runesson R, Tuvhag R, et al. Intraarticularcorticosteroid injection: pain relief in osteoarthritis of thehip? J Rheumatol 2004;31:2265.

19. Plant MJ, Borg AA, Dziedzic K, et al. Radiographic patternsand response to corticosteroid hip injection. Ann RheumDis 1997;56:476.