Secondary Logo

Journal Logo

Does Pain Predict Outcome in Hips with Osteonecrosis?

Belmar, Carlos, J; Steinberg, Marvin, E; Hartman-Sloan, Karen, M

Clinical Orthopaedics and Related Research: August 2004 - Volume 425 - Issue - p 158-162
SECTION II: ORIGINAL ARTICLES: Hip
Free
SDC

It generally is accepted that without specific treatment 70–80% of hips with clinically diagnosed osteonecrosis will progress to collapse. However, there are conflicting reports regarding the relationship between pain and outcome before femoral head collapse. Some surgeons are reluctant to operate on patients with asymptomatic or minimally symptomatic hips, assuming that these patients have a better prognosis than patients with pain. This study reviewed the outcome of 328 hips in 235 patients with nontraumatic osteonecrosis, all treated with core decompression and grafting. The preoperative stage, the extent of necrosis, and the Harris pain scores were correlated with the clinical and radiographic outcomes. Mean followup was 46 months. Patients with hips treated surgically did better as a group than patients with hips treated without surgery. A direct correlation was found between outcome and the stage and size of the necrotic lesion. Hips that had femoral head collapse were more painful than hips that did not have collapse and had a poorer outcome. Before collapse, outcome was correlated with the size of the necrotic lesion but there was no correlation with the preoperative pain level. These findings, although limited to patients with hips which had core decompression and grafting, support the observations of investigators who reported that most asymptomatic hips with osteonecrosis would progress without specific treatment. They also may apply to hips which have other forms of prophylactic treatment. Although several factors must be considered in determining the optimum treatment of hips with early stages of osteonecrosis, prophylactic treatment should not be withheld specifically because of the absence or paucity of pain.

From the Department of Orthopaedic Surgery, University of Pennsylvania School of Medicine, Philadelphia, PA.

Received: April 2, 2003

Revised: August 11, 2003; November 11, 2003

Accepted: January 16, 2004

Correspondence to: Marvin E. Steinberg, MD, Department of Orthopaedic Surgery, Hospital of the University of Pennsylvania, 3400 Spruce Street, Philadelphia, PA 19104. Phone: 215-662-3340; Fax: 215-349-5928; E-mail: marvin.steinberg@uphs.upenn.edu.

The etiology of hip pain in patients with osteonecrosis (ON) of the femoral head has not been definitively established and may be multifactorial. It has been documented that without specific treatment 70–80% of hips with clinically established ON progress to femoral head collapse.1,2,8,11,13,17–19 Numerous surgeons, therefore, recommend early diagnosis and treatment to retard progression and improve the outcome. However, there is a difference of opinion as to the relationship between pain and outcome in hips diagnosed before femoral head collapse. Accordingly, some are reluctant to do prophylactic procedures in a patient who is asymptomatic or minimally symptomatic.4,18,20 This is based largely on the assumption that these patients have a better prognosis than patients with symptoms. This may be correct for patients with asymptomatic hips with small lesions, in which the prognosis is relatively good and for patients with painful hips with collapse, in which the prognosis is poor. However, there is no documentation that pain is a valid predictor of outcome in hips diagnosed before femoral head collapse. This study was done to evaluate the relationship between pain and outcome in hips with osteonecrosis.

Back to Top | Article Outline

MATERIALS AND METHODS

The study included 328 hips in 235 patients with nontraumatic ON of the femoral head. Initially, patients were evaluated clinically and Harris hip scores15 were determined by the senior author (MES). All patients had AP and lateral radiographs of both hips. Magnetic resonance imaging (MRI) was done on patients with normal radiographs. All patients in this series had core decompression and grafting as described previously.17 Patients were followed up every 3 months for the first year, every 6 months for the second year, and yearly thereafter. Hip scores were determined and radiographs were repeated at these intervals. The stage and extent of necrosis were determined preoperatively and postoperatively by two examiners using the University of Pennsylvania Staging System (Table 1).15,16 For the purpose of calculation, each stage was assigned a specific numeric value with the least affected hips (Stage 0) given 1 point, and the most affected hips (Stage VI) given 21 points.14,16,17 Postoperative complications were noted and hips which required subsequent total hip replacement (THR) were recorded.

Table 1

Table 1

Hips were graded by radiographic stage and lesion size at the time of surgery. In each stage hips additionally were subdivided into five groups depending on the Harris pain score: no pain (44 points); minimal pain (43–40 points); mild pain (39–30 points); moderate pain (29–20 points); and severe pain (19–0 points).

The results of the core decompression and grafting were determined by calculating the change in total Harris hip score, the degree of radiographic improvement or progression, and the need for THR. Five patients (eight hips) died and 13 patients (18 hips) could not be located for a minimum 2-year followup. This left 302 hips in 217 patients available for followup. The mean followup was 46 months (range, 1–156 months).

An analysis of variance (ANOVA) was done to evaluate the prevalence of THR as related to the size of the necrotic lesion.

Back to Top | Article Outline

RESULTS

At the time of surgery there were 64 hips in Stage I, 137 in Stage II, 14 in Stage III, and 87 in Stage IV. Of these, 45 had no pain, 59 had minimal pain, 64 had mild pain, 98 had moderate pain, and 36 had severe pain. In the 217 patients who completed the study, 113 of 302 hips (37%) required THR.

No differences were seen in the percentage of THRs required among Stages I, II, and III, but there was a higher percentage of THRs in Stage IV than in these earlier stages (Fig 1). This was anticipated because hips in Stage IV, by definition, already had femoral head collapse before surgery. In Stage IV, 78 hips (90%) had significant pain initially and 42 hips (48%) eventually required THR. It was impossible to divide these hips into five separate pain groups, and no meaningful conclusions could be drawn regarding the specific relationship between pain and outcome in Stage IV. The Stage III hips were those in which there was a subchondral fracture or crescent sign without femoral head flattening. There were only 14 hips in this stage and because of the small number it was not possible to correlate outcome with pain. Therefore the definitive evaluations focused primarily on hips in Stages I and II, before femoral head collapse.

Fig 1.

Fig 1.

Hips with small lesions in Stages I and II had a lower prevalence (p = 0.048) of THRs than hips with medium or large lesions, however there was no difference between hips with medium and hips with large areas of necrosis (Fig 2). In Stages I and II there was no difference in the amount of pain recorded for hips with small, medium, or large lesions (Fig 3). There also was no correlation between the amount of pain present initially in Stages I and II and the eventual need for THR (Fig 4) or the radiographic progression. This was true when hips were evaluated by overall stage and when they were evaluated specifically by the size of the necrotic lesion (Fig 5).

Fig 2.

Fig 2.

Fig 3.

Fig 3.

Fig 4.

Fig 4.

Fig 5.

Fig 5.

In Stage I, hips with little or no pain (40–44 points) showed mean radiographic progression of the necrotic lesion by 7.6 points on a scale of 21, whereas hips with mild to severe pain (0–39 points) progressed 6.6 points. In Stage II, patients with hips with little or no pain progressed 9.2 points and those with mild to severe pain progressed 11.0 points.

In Stages I and II, hips which did not require THR had a mean gain of 24 points in their Harris hip score between the preoperative and final postoperative evaluations, whereas hips that required THR lost a mean of 31 points. Hips which initially had little or no pain lost a mean of 22 points, whereas hips that initially had mild to severe pain gained a mean of 6 points. This implies that patients with relatively asymptomatic hips actually had a worse clinical outcome than patients with painful hips, however this observation can be misleading. Because the Harris hip score is weighted heavily toward pain (44 of 100 points),5 patients with asymptomatic hips with a high initial score could gain few if any points if they did well clinically, yet could lose a significant number of points if the necrosis progressed. Conversely, patients with painful hips with a low initial score could lose few points if they had additional deterioration, yet could increase their pain score considerably if they had clinical improvement. Therefore the change in the Harris hip score is not a valid method to compare improvement or deterioration specifically between painful hips and hips which were asymptomatic or minimally symptomatic.

Back to Top | Article Outline

DISCUSSION

It generally is accepted that most hips clinically diagnosed with ON will progress to collapse if untreated.1,2,11,13,17–19 Although there are differing opinions as to which surgical procedure is best, many agree that some form of early surgical intervention before femoral head collapse generally will yield better results than symptomatic treatment.2,11,15–19 Many surgeons, however, are reluctant to surgically treat hips which are asymptomatic or have minimal symptoms.4,18,19 This is based in part on the assumption that the prognosis for these patients, without treatment, is better than for patients with hip pain. However, this has not been documented. Sporadic reports have appeared in the literature, but most are retrospective studies involving small numbers of patients and their conclusions differ substantially.3,4,6,8–10,20,21

Our study was done to determine the relationship between pain and outcome in hips with ON. It was prospective and included 328 hips, 201 of which were Stages I or II, which were evaluated clinically and radiographically. All hips had core decompression and grafting done by one surgeon, following a standard regimen. It did not include hips which had primary THRs, nor did we include untreated hips in this study because of the documented history of progression without treatment. Such cases have been described in the literature cited. The overall outcome for hips treated by core decompression, with or without supplemental bone grafting, has been shown to be better than for hips treated nonoperatively at our institution15,17 and by other investigators.1,2,11,13,18,19

The literature reflects differing opinions regarding the relationship between pain and outcome in patients with ON. In 1980 Lee et al10 reported on 11 untreated, asymptomatic hips with ON. All eventually showed radiographic progression. In 1993 Bradway and Morrey3 described 15 asymptomatic and radiographically normal hips in patients with contralateral ON. They found that all of these hips eventually collapsed at a mean of 23 months. However, they did not discuss their criteria for excluding numerous patients from this study. Takatori et al21 evaluated 32 radiographically normal, asymptomatic hips in 25 patients treated nonoperatively. The diagnosis of ON was made using MRI. They found no evidence of femoral head collapse in 15 cases with small lesions. However, 14 of 17 hips (82%) with moderate to large lesions collapsed in a mean of 15 months. Therefore they found a clear correlation between outcome and lesion size but no correlation with pain. Koo et al8 observed 15 patients with asymptomatic ON, treated nonoperatively. Eleven (73%) had femoral head collapse. Jergesen and Khan6 reported that of 75 patients with nontraumatic ON 46 (61%) initially had pain in only one hip. Twenty-two of the asymptomatic hips had radiographic evidence of ON. Three of these had core decompressions, whereas 19 were untreated. Fourteen of these 19 untreated hips (74%) eventually became painful. Davidson et al4 observed 133 patients with known, symptomatic ON in one hip in which the opposite hip was asymptomatic. Of 77 asymptomatic hips with no radiographic or MRI evidence of ON, only six (7.8%) later had ON develop. Fifty-six asymptomatic hips did have ON diagnosed initially by plain radiographs or MRI. Forty one (73%) of these progressed and became symptomatic at an average of 24 months from the initial observation: two of seven (29%) with Stage I lesions, 34 of 43 (79%) with Stage II lesions, and five of six (83%) with Stage III lesions. Davidson et al4 concluded that more than 79% of patients who present with asymptomatic Stages II and III ON will progress to symptomatic disease and might be candidates for early surgical intervention, although this was not their standard practice at the time.

Not all authors share these opinions. Kopecky et al9 identified 25 hips with asymptomatic ON, diagnosed by MRI, in 104 renal transplant recipients. Pain and radiographic changes developed in seven hips (28%) and one hip became symptomatic despite a normal radiograph. In seven hips the MRI showed a decrease in lesion size and in six of these the appearance of the hips on MRI scans eventually became normal. Sugioka et al20 observed that in some conservatively treated hips with asymptomatic ON, they were unable to detect progression, and that occasionally in hips with small lesions the necrotic area decreased. Specific numbers were not cited in their study. Mulliken et al12 diagnosed ON in 15 hips (10 patients) of 132 renal transplant recipients evaluated by MRI. Eleven hips were preradiographic and asymptomatic. Only one of these eventually progressed. Four hips were symptomatic and radiographically positive for ON, and an additional seven patients with previously diagnosed ON were excluded from the study.

Therefore in five of the studies which specifically evaluated asymptomatic hips with osteonecrosis,3,4,6,8,10 73–100% progressed without treatment. Takatori et al21 found, as did we, that small lesions had a better prognosis than moderate to large lesions. No collapse was seen in 15 hips with small lesions, whereas 14 of 17 with moderate to large lesions (82%) collapsed. In three studies9,12,20 progression was noted in only 9–28% of hips.

These findings compare closely with the work of previous investigators who have reported overall progression in 70–80% of untreated hips with clinically established ON, many of which were painful.1,2,11,13,17–19 This stands in contrast to the 37% prevalence of THRs for hips treated surgically in the current study and by other investigators.2,11,15,17–19 Although the overall prevalence of progression without specific treatment is higher than after surgical treatment, in neither group was there evidence that asymptomatic or minimally symptomatic hips had a better prognosis than hips with pain, if lesion size and stage were considered.

In the current study we found that hips with femoral head collapse were more painful and had a worse prognosis than hips without collapse. Before collapse, outcome was correlated with lesion size, thereby confirming previous observations.7,14,16,17,19 No correlation was found between lesion size and pain, nor was there a correlation between pain and outcome as determined by radiographic progression and the need for THR. Despite the beneficial effects of core decompression, numerous relatively asymptomatic hips deteriorated.

Our observations, although limited to hips which had core decompression and grafting, support previous reports which found that pain did not predict the outcome of untreated hips with ON, seen before femoral head collapse. These conclusions also might be applicable to hips which have forms of prophylactic treatment other than core decompression. Asymptomatic or minimally symptomatic hips do not have a better prognosis than hips with significant pain if stage and lesion size are accounted for. Although many factors must be considered in determining the optimum treatment of patients with ON, early prophylactic surgery should not be withheld solely because of the absence or paucity of pain.

Back to Top | Article Outline

References

1. Aaron RK, Lennox D, Stulberg BN. The Natural History of Osteonecrosis of the Femoral Head and Risk Factors for Rapid Progression. In Urbaniak JR, Jones JP (eds). Osteonecrosis: Etiology, Diagnosis and Treatment. Rosemont, IL, American Academy of Orthopaedic Surgeons 261–265, 1997.
2. Arlet J. Nontraumatic avascular necrosis of the femoral head: Past, present and future. Clin Orthop. 1992;277:12–21.
3. Bradway J, Morrey BF. The natural history of the silent hip in bilateral atraumatic osteonecrosis. J Arthroplasty. 1993;8:383–387.
4. Davidson JJ, Coogan PG, Gunneson EE, et al. The Asymptomatic Contralateral Hip in Osteonecrosis of the Femoral Head. In Urbaniak JR, Jones JP (eds). Osteonecrosis: Etiology, Diagnosis and Treatment. Rosemont, IL, American Academy of Orthopaedic Surgeons 231–240, 1997.
5. Harris WH. Traumatic arthritis of the hip after dislocation and acetabular fractures: Treatment by mold arthroplasty. J Bone Joint Surg. 1969;51A:737–755.
6. Jergesen HE, Khan AS. The natural history of untreated asymptomatic hips in patients who have nontraumatic osteonecrosis. J Bone Joint Surg. 1997;79A:359–363.
7. Koo K-H. A new Method of Quantifying the Extent of Necrosis and Predicting the fate of Osteonecrosis of the Femoral Head. In Urbaniak JR, Jones JP (eds). Osteonecrosis: Etiology, Diagnosis and Treatment. Rosemont, IL, American Academy of Orthopaedic Surgeons 225–230, 1997.
8. Koo K-H, Kim R, Ko G-H, et al. Preventing collapse in early osteonecrosis of the femoral head. J Bone Joint Surg. 1995;77B:870–874.
9. Kopecky KK, Braunstein EM, Brandt KD, et al. Apparent avascular necrosis of the hip: Appearance and spontaneous resolution of MR findings in renal allograft recipients. Radiology. 1991;179:523–527.
10. Lee CK, Hansen HT, Weiss AB. The “silent hip” of idiopathic ischemic necrosis of the femoral head in adults. J Bone Joint Surg. 1980;62A:795–800.
11. Mont MA, Hungerford DS. Current concepts review: Nontraumatic avascular necrosis of the femoral head. J Bone Joint Surg. 1995;77A:459–474.
12. Mulliken BD, Renfrew DL, Brand RA, et al. Prevalence of previously undetected osteonecrosis of the femoral head in renal transplant recipients. Radiology. 1994;192:831–834.
13. Musso ES, Mitchell SN, Schink-Ascani M, et al. Results of conservative management of osteonecrosis of the femoral head: A retrospective review. Clin Orthop. 1986;207:209–215.
14. Steinberg ME, Bands RE, Parry S, et al. Does lesion size affect outcome in avascular necrosis? Clin Orthop. 1999;367:262–271.
15. Steinberg ME, Brighton CT, Corces A, et al. Osteonecrosis of the femoral head: Results of core decompression and grafting with and without electrical stimulation. Clin Orthop. 1989;249:199–208.
16. Steinberg ME, Hayken GD, Steinberg DR. A quantitative system for staging avascular necrosis. J Bone Joint Surg. 1995;77B:34–41.
17. Steinberg ME, Larcom PG, Strafford B, et al. Core decompression with bone grafting for osteonecrosis of the femoral head. Clin Orthop. 2001;386:71–78.
18. Steinberg ME, Mont MA. Osteonecrosis. In Chapman MW (ed). Chapman’s Orthopaedic Surgery. Ed 3. Philadelphia, Lippincott Williams & Wilkins 3263–3308, 2001.
19. Stulberg BN, Bauer TW, Belhobek GH. Making core decompression work. Clin Orthop. 1990;261:186–195.
20. Sugioka Y, Hotokebuchi T, Tsutsui H. Transtrochanteric anterior rotational osteotomy for idiopathic and steroid-induced necrosis of the femoral head: Indications and long-term results. Clin Orthop. 1992;277:111–119.
21. Takatori Y, Kokubo T, Ninomiya S, et al. Avascular necrosis of the femoral head: Natural history and magnetic resonance imaging. J Bone Joint Surg. 1993;75B:217–221.
© 2004 Lippincott Williams & Wilkins, Inc.