The Journal of Arthroplasty Vol. 8 No. 4 1993

The Natural History of the Silent Hip in Bilateral Atraumat ic Osteonecrosis

J o h n K. B r a d w a y , M D , * a n d B e r n a r d F. M o r r e y , M D

Abstract: The natural history of presymptomatic ischemic necrosis of the femoral head was studied in the contralateral hip of 15 patients undergoing total hip arthroplasty for avascular necrosis between 1979 and 1984. These 15 hips were retrieved from among 416 patients undergoing total arthroplasty during this time period. Presyrnptomatic avascular necrosis was defined according to three strict criteria that allowed the hips to be studied from an index entry point (no evidence of avascular necrosis). Follow- up examinations of each of these 15 hips from this entry point revealed that all hips eventually collapsed at a mean of 23 months after index entry point (range, 3-66 months). These data indic&e that the "silent hip" is at significant risk of developing avascular necrosis and, if it becomes involved, it progresses to collapse in a high per- centage of patients in a relatively short time. There was no correlation between under- lying etiology and the time to collapse. K e y words: hip, osteonecrosis, natural history, ischemic necrosis.

Ischemic necrosis of the femoral head is a well- known sequelae to femoral neck fractures and hip dislocations. Additionally, ischemic necrosis of the femoral head is found in association with several atraumatic conditions, such as collagen vascular dis- ease, corticosteroid use, hemoglobinopathies, alco- hol abuse, and caisson disease, t'3 In a significant per- centage of cases, no known risk factors for the disease a r e found and the disease is considered idiopathic?

For clinicians, the greatest concerns of ischemic necrosis of the femoral head are generally associated with the nontraumatic etiologies of the disease. Two of the larger problems are (1) the high propensity of nontraumatic ischemic necrosis of the femoral head to affect both hips, reportedly~n up to 80% of cases and (2) the largely unknown question of the natural history of the untreated disease. ~'6 There have been significant contributions to the literature regarding diagnostic techniques and treatment of ischemic he-

From the Department of Orthopedics, Mayo Clin& and Mayo Founda. lion, Rochester, Mhlnesota.

*Currently at Practice Limited to Orthopedic Surgery, Phoenix, Arizona.

Reprint requests: Bernard F. Morrey, MD, Mayo Clinic, 200 First Street SW, Rochester, MN 55905.

crosis of tile femor/tl head. 2-4 Ilowever, the impor- tance of treatment regimens and early diagnosis can only be fully realized when the natural history of the untreated disease is known. To date, the natural history of ischemic necrosis of the femoral head has not been well delineated nor adequately described in the literature. This is because the diagnosis of the disease occurs at the time the patient presents with symptoms, usually late into the course of radio- graphic changes of the involved hip. Thus, it is diffi- cult to define an adequate starting point from which to track the disease forward to the time of collapse of the femoral head and subsequent arthritic changes of the involved hip. Further, many patients identified with ischemic necrosis of the femoral head have in- vasive diagnostic procedures or some form of opera- tive or nonoperat ive treatment sometime during the course of the disease, creating an additional modifier of the actual natural history of ischemic necrosis of the femoral head. Bailey and Miller have stated that " the natural history of osteonecrosis is known only after subchondral collapse has occurred. ''~

Many authors speculate the natural history of is- chemic necrosis of the femoral head to be progressive without treatment. 3,5.6 While this is generally held to

383

384 The Journal of Arthroplasty Vol. 8 No. 4 August 1993

be true, there is little documentation in the literature supporting this claim. Merle D'Aubigne et al. stated that 75% of 90 hips followed had collapsed by 3 years. 5 However, this time span was measured from the time symptoms began, and it is not clear as to what the radiographic appearance of the hips were at entry, nor if the patients underwent any treatment during the follow-up period. Steinberg et al. has pub- lished that progression occurred in 92% of 48 hips that undenvent nonoperative management, but the specific time to collapse, state of tile hip at the time management began, and analysis of the result by treatment (full, partial, or non-weight bearing) was ]lot given. 6

In this study we have followed nontraumatic is- chemic necrosis of the femoral head in a selected group of untreated patients known to develop os- teonecrosis eventually to document the true progres- sion of the disease over time. To our knowledge, this is the only study to address specifically the natural history of presymptomatic, untreated ischemic ne- crosis of the femoral head as its primary focus.

Materials and Methods

The records of 416 consecutive patients, treated between 1979 and 1984 at the Mayo Clinic (Roches- ter, MN), with a diagnosis of nontraumatic ischemic necrosis of the femoral head who underwent total hip arthroplasty on at least one hip were carefully reviewed. From this group of 416 patients, each record was studied to identify those in whom the opposite hip was asymptomatic with neither clinical nor radiographic evidence of the disease. The three criteria used for inclusion in the study and for this entry point were: (1) no evidence in the record of any symptom related to the hip in question; (2) a plain anteroposterior pelvis radiograph that was in- terpreted as negative for any conclusive sign of ne- crosis; and (3) a negative technetium-99m bone scan, if available. The date that a patient satisfied these criteria was labeled as the "index entry point." In addition, to be included in the study, the patient received postoperative"hip arthroplasty pelvis radio- graphs at 3, 12, and 24 months as follow-up data of the opposite replaced hip (nonstudy hip). Radio- graphs at these intervals are part of the routine fol- low-up examinations of total hip arthroplasty pa- tients at the Mayo Clinic, and this allowed follow- up evaluation of the silent hip in these patients as well. The endpoint of the silent hip study in each individual patient was considered to be collapse of the femoral head and/or when definitive treatment

was necessitated. In addition, at no time during the course of the investigation did the studied hip receive treatment other than crutch walking for up to 3 months after the replaced hip and analgesics for pain. In this way, tile time period from a preradiographic, asymptomatic entry point to an uneventftfl course or to collapse of the femoral head and/or arthritic changes in the tmtreated hip could be determined, and the natural history of the disease defined.

Fifteen hips in 15 patients satisfied the strict criteria and were included for study. There were eight men and seven women with a mean age of 58 years (range, 30-74 years). Tile right hip was studied in seven patients and the left hip in eight. The etiology of ischemic necrosis of the femoral head was steroid use in seven patients, idiopathic in two, and alcohol abuse in six. Four patients (4 hips) had negative tech- netium-99m bone scans at the index entry point. While all patients studied had a normal interpreta- tion of an anteroposterior pelvis radiograph, three patients also had oblique radiographs available at tile index entry point indicating no evidence of ischemic necrosis of the femoral head.

Seven hips had subsequent surgical confirmation of ischemic necrosis of the femoral head by gross examination of the femoral head at the time of a subsequent hip arthroplasty. Three hips had both surgical and pathological confirmation of ischemic necrosis of the femoral head and three hips had path- ological confirmation without comment in the surgi- cal record on the appearance of the femoral head at surgery. The two remaining hips had clinical and radiographic evidence of iscl~mic necrosis of the femoral head, but both patients died before undergo- ing total hip arthroplasty. By definition of the study group, both patients previously had total hip arthroplasty in the opposite hip for documented is- chemic necrosis of the femoral head.

Results

All 15 hips entered into the study undenvent pro- gression to eventual collapse of the femoral head at a mean of 23 months (range, 3-66 months) from the index entry point (Table 1). In the majority of cases, total hip arthroplasty was required for relief of pain and was performed shortly after evidence of collapse of the hip. The etiology of ischemic necrosis of the femoral head did not have a significant bearing on the mean time to collapse; the mean times to collapse of tile femoral head for steroid, idiopathic, and alcohol abuse etiologies were 19, 43, and 20 months, respectively.

Atraumatic Ischemic Necrosis �9 Bradway and Morrey 385

Table 1. Characteristics of Patients with Presymptomatic Avascular Necrosis

Index Entry to Collapse Patient No. Age (years) Sex Study tlip Etiology (mo)

! 64 Female Right Steroids l 0.0 2 48 Male Left Alcohol 9.5 3 62 Male Left Steroids 66.0 4 58 Female Right Alcohol 3.0 5 46 N[ale Right Steroids 9.0 6 60 Female Left Alcohol 50.0 7 68 Female Left Idiopathic 25.0 8 38 Male Right Steroids 6.0 9 56 Male Right Alcohol 4.0

10 72 Male Left Steroids 5.0 11 70 Female Left Stcroids 14.0 12 61 Male Right Alcohol 19.0 13 63 Female Left Alcohol 33.0 14 74 Female Left Idiopathic 6 !.0 15 30 Male Right Steroids 22.0

The period of time from index entry point to col- lapse of tile hip was quite variable. Thirty-three per- cent of hips progressed to collapse between 1 and 3 years from the index entry point (Fig. 1). tIowever, seven hips (47%) went on to collapse in less than 1 year, while three hips (20%) collapsed at 6 years or later.

C a s e Reports

Case One

A 48-year-old man with a long history of alcohol abuse underwent a right total hip arthroplasty in July 1981 for avascular necrosis. Anteroposterior pelvis radiographs plus oblique and lateral radio- graphs taken prior to the time of tile right total hip arthroplasty were negative for any pathology on the left hip. A bone scan was also performed at that time

and found to be negative. Over the ensuing months the patient became symptomatic on the left side. Re- peat plain films in November 1981 revealed ad- vanced changes of ischemic necrosis of the femoral head on the left side. Computed tomograms and bone scans also were performed, and both of these studies were positive, with the tomograms showing collapse of the femoral head. The patient was symp- tomatic enough to require a total hip arthroplasty in February 1982. Although tile femoral head was not submitted to pathology for histological diagnosis, the surgical record commented oil tile deformity and col- lapse of the femoral head at the time of the total hip arthroplasty.

Case Two

A 46-year-old man was diagnosed with Addison's disease. He was placed on steroids for treatment of his problem for several years, tie underwent a total

Fig. 1. Cumulative percentage of patients with avascular ne- crosis demonstrating collapse of the asymptomatic hip.

100

80

60 E o

4o (n

-~ 2o 0

I I I I I I I 1 2 3 4 5 6

Time, yr

386 TheJournal of Arthroplasty Vol. 8 No. 4 August 1993

hip arthroplasty in March 1979, and ischemic necro- sis of the femoral head was confirmed on pathologic diagnosis. Routine follow-up examinations with an anteroposterior pelvis radiograph of the total hip was negative for any involvement of the right hip in Feb- mary 1980. A bone scan was not performed. The patient developed symptoms of pain in the right hip, and an additional anteroposterior pelvis radiograph was taken in November 1980 that showed advanced signs of avascular necrosis and collapse. Tile patient underwent a total hip arthroplasty on the right side 1 week later. Comment in the surgical record revealed collapse and clinical findings of ischemic necrosis of the femoral head, and this was confirmed with pathologic and histologic diagnosis.

Discussion

The results of this study are striking. All hips satis- fying the inclusion criteria for this study went on to collapse from an apparent preradiographic and asymptomatic entry point. The majority collapsed in less than 3 years.

The mean progression time from the defined index entry point to eventual collapse of the hip was very rapid, and somewhat surprising, since there were no significant differences among the various etiologies for the ischemic necrosis. The progression of disease in this study is not only more rapid than previously thought, 5 but in fact, the time from entry to collapse may be an overestimation in many cases because the index entry point was arbitrarily determined by satis- faction of the inclusion criteria. In some cases, the

�9 index entry point was several months or years before the diagnosis became clinically or radiographically apparent, simply because the most recent negative radiograph prior to the development of the disease was often taken some time earlier than tile disease manifested itself to the clinician.

Clearly, this study employed a selected-sample of patients since the silent hip was diagnosed in patients known to have developed severe enough necrosis to justify hip arthroplasty on the opposite side. None- theless, there were no other means available to iden- tify the presymptomatic joint during this particular time period of study. We recognize that the an te ro - posterior pelvis radiograph is not the most sensitive radiographic view to visualize early collapse. I t is generally accepted that a frog lateral view is more sensitive for the diagnosis of ischemic necrosis of the femoral head. However, we felt that the anteropos- terior pelvis radiographs did represent useful criteria for the clinicians sificc precollapse radiographic changes of the ischemic necrosis of the femoral head

do eventually become visible on the routine antero- posterior pelvis view. Of interest is that in four cases, t echnet ium-99mm bone scans were performed at about the same time as the negative plain radio- graphs used for the index entry point. In all four cases, the bone scans were interpreted as negative on the uninvolved side despite the fact that the ra- diologist was aware of tile presumed diagnosis. In at least one recent study, technet ium-99m bone scans have been found to have 100% sensitivity for is- chemic necrosis of tile Femoral head. 3 Hence it is likely that the disease had not yet affected those hips with a negative scan.

Finally, it is entirely possible that patients may also have had histologic evidence of ischemic necrosis of the femoral head even at the time of our index entry point, even though it was unapparent by all available information. Since our institution has not, until re- cently, utilized functional exploration of bone (core decompression) in confirming tile diagnosis of is- chemic necrosis of the femoral head as described by Ficat, 2 and later Hungerford, 4 such information was not available to possibly exclude some patients from this study. However, functional exploration of bone adds an undesirable variable in determining the true natural history of the untreated disease because it is an invasive procedure. Magnetic resonance imaging was not available during this study period, but should allow for further study and delineation of the natural history of the femoral head, possibly on a prospective study basis.

Regardless of the possible selection bias of this study, we feel that these findings are of some value to the clinician treating patients with ischemic necro- sis of tile femoral head. To date, there are no data available in tile literature that offer any basis or allow prediction of the rate of progression of the asymp- tomatic hip in a patient with known capital femoral necrosis. From a preradiographic and presymptom- atic index point, tile patient with untreated ischemic necrosis of the femoral head in this study progressed to collapse in every case, with a mean time to collapse of just under 2 years. While this finding cannot be extrapolated to include all such patients, and while each hip may have a variable course to collapse, the data may serve as a valuable prognostic guideline for some patients, and possibly as a basis for the measure of success or failure of different treatment modalities for this complex disease.

References

]. Bailey JP, Miller ME: Osteonecrosis of tile femoral head: treatment before collapse. A critical review. Adv Orthop Surg 55, 1988

Atraumatic Ischemic Necrosis �9 Bradway and Morrey 387

2. Ficat RP: Idiopathic bone necrosis of the femoral head: early diagnosis and treatment. J Bone Joint Surg 67B: 3, 1985

3. Hopson CN, Siverhus SW: Ischemic necrosis of the fem- oral head: treatment by core decompression. J Bone Joint Surg 70A:1048, 1988

4. Hungerford DS: Bone marrow pressure, venography, and core decompression in ischemic necrosis of the fem- oral head. p. 218. In The hip. Proceedings of the Sew

enth Open Scientific Meeting of the ttip Society. CV Mosby, St. Louis, 1979

5. Merle D'Aubigne R, Postel M, Mazabraud A et al: Idio- pathic necrosis of the femoral head in adults. J Bone Joint Surg 47B:612, 1965

6. Steinberg ME, Brighton CT, Steinberg DR et al: Treat- ment of avascular necrosis of the femoral head by a combination of bone grafting, decompression, and elec- trical stimulation. Clin Orthop 186:137, 1984