DIFFERENTIAL DIAGNOSIS

The differential diagnosis of the painful THA can be organized into two divisions: causes that are intrinsic to the hip and causes that are extrinsic to the hip ( Table 30-1 ). This comprehensive list of conditions can be used as a starting point for the workup.

HISTORY

A detailed history and careful examination are important first steps in the evaluation of the painful THA. This information can be used to narrow the differential diagnosis considerably, leading to a more efficient and cost-effective workup. The history should focus on the various aspects of the patients’ pain complaint: its onset, location, timing, and palliative and provocative factors.

Pain that continues to be exactly the same after surgery as before the surgery suggests a possible extrinsic cause for the hip pain that was not addressed by the THA. On the other hand, pain that is different from the preoperative pain and that starts immediately after the surgery suggests an early postoperative complication, such as unrecognized intraoperative fracture, dislocation, poor fixation leading to early loosening of the implant, acute infection, or hematoma. Pain that starts to occur much later, after a long pain-free interval, is more consistent with aseptic loosening, osteolysis, periprosthetic fracture, late infection, bursitis, or tendonitis. Late pain can also be due to an extrinsic cause, exacerbated by the increased activity that a successful THA typically allows.

The location of the pain can be very helpful, although some patients find it hard to precisely localize their symptoms. Pain felt in the groin or deep buttock is consistent with problems with the acetabular component. However, groin pain is also associated with inguinal, femoral, or obturator hernia ; iliopsoas tendonitis due to acetabular retroversion and impingement ; and a number of genitourinary and gynecologic diseases. Moreover, pain in the buttock could be from referred low back pain, sacroiliac disease, or piriformis syndrome.

Pain over the greater trochanter is frequently due to bursitis, sometimes related to underlying sutures or wires, but can also be from trochanteric fracture or nonunion.

Anterior thigh pain is associated with femoral stem loosening, as well as mismatch of stiffness between the host bone and femoral stem. Thigh pain can also be due to neurogenic referred pain, especially posterolateral thigh pain that is numb and tingling in character and radiates below the knee in a dermatomal distribution. This is most likely from lumbar radiculopathy, but could also be from iatrogenic injury to the sciatic nerve. Finally, if the pain is in the calf, consider neurogenic or vascular claudication.

In regard to timing, pain at rest or at night is associated with infection or malignancy. In particular, constant mild pain after THA is suggestive of infection. Pain that is aggravated by activity and is relieved with rest could be from tendonitis or a loose component. In particular, pain that follows a triphasic pattern of initial pain on starting activity, followed by a lessening of pain, then increasing with prolonged activity, is consistent with a loose component in its early stages. Pain that occurs only in certain positions suggests subluxation or dislocation.

The patient should also be asked about any possible precipitating events. Pain beginning with a fall may be from a fracture or traumatic loosening. A recent distant infection, a dental, gastrointestinal, or genitourinary procedure, a systemic illness, and wound healing problems, including postoperative hematoma or persistent drainage, raise the possibility of infected THA. Past medical history can also reveal factors that increase the chance of infection, such as diabetes, obesity, chronic inflammatory conditions, immunocompromised status, and previous hip surgeries.

PHYSICAL EXAMINATION

The physical examination should include a thorough assessment of both hips, both knees, and the spine, to assess both the painful hip as well as other potential sources of referred pain. Starting in the standing position, one should inspect the spinal alignment for scoliosis, kyphosis, lordosis, or pelvic obliquity, and look for gluteal, hamstring, or quadriceps muscle atrophy. Single-leg stance should be tested for several seconds to see if Trendelenburg sign will turn positive; this sign is consistent with abductor weakness, trochanteric nonunion, gluteal nerve injury, adduction deformity, ankylosis, as well as THA failure. Standing leg length measurements should be performed with blocks to determine true versus apparent leg length discrepancy. A gait examination is then done, both with and without walking aids, if applicable, to check for antalgic gait or Trendelenburg gait and to assess balance and mobility.

Moving then to a supine position, inspection should be done to check for distal muscle atrophy or asymmetry, overall limb alignment, and supine leg lengths. Skin examination is done to look for dermatitis, cellulitis, abscesses, ulceration, thin skin, or chronic or active sinuses. Palpation about the hip may reveal tenderness over the trochanter in trochanteric bursitis or nonunion, over the sacroiliac joint for sacroiliitis, over the ischium for hamstring tendonitis, over the pubic rami for pelvic stress fracture, over the gluteus maximum insertion site, or over the spinous processes. Palpation can also reveal fascial defects and masses due to hernia, heterotopic ossification, or ectopic cement.

Range of motion of the hip is tested through both active and passive arcs, and any painful response is noted. Pain only at the extremes of motion can be seen with loosening, whereas pain throughout the arc of motion can be seen with active synovitis or infection. Stability examination must be performed very carefully lest frank dislocation occur in the office setting. Strength examination may reveal weakness in pure flexion, which could occur in the setting of iliopsoas tendonitis or previous psoas tendon release, or weakness in abduction, from abductor tendonitis or avulsion of the gluteus medius tendon.

Because one of the more common extrinsic differential causes of hip pain is referred pain from a primary pathologic process of the back, the back and neurologic examination is an important component of the workup for painful THA. Careful attention should be given to the sensory, motor, and reflex function of the femoral nerve, the obturator nerve, and both common peroneal and tibial branches of the sciatic nerve. Stretch tests can be performed on the sciatic and femoral nerves. Attention should be given to possible dermatomal or myotomal patterns of dysfunction. Finally, the vascular and lymphatic systems should also be evaluated by noting distal pulses, atrophic skin changes, or peripheral edema.

LABORATORY TESTS

Peripheral white blood cell count (WBC), erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) levels are commonly used in the workup of the possibly infected THA. At this time, it is generally acknowledged that the WBC has little utility in this setting, owing to its low sensitivity. Canner and coworkers reported that in only 15% of 52 known cases with infected THA was there an elevated WBC. In addition, Spangehl and associates found in a prospective study of 178 patients having revision THA that the WBC had only sensitivity of 20% for infection.

The ESR, on the other hand, is a very sensitive marker for inflammation. It increases after THA and returns to less than 20 mm/hr within 6 months. Forster and Crawford reported that their group of patients with established infected THA started with a mean ESR of 60 mm/hr, which then dropped to below 20 mm/hr after the infection was eradicated. Sensitivity of the ESR to infection has been reported at between 78% and 100%. However, the ESR is not very specific and can be elevated in rheumatoid arthritis, connective tissue disorders, malignancy, and pregnancy.

More helpful even is the CRP, because it returns to normal within 3 weeks postoperatively. Sanzen and Carlsson found that only 1 of 23 infected hips, but all 33 infected hips, had a CRP less than 20 mg/L. Like the ESR, it is a very sensitive marker for inflammation, but with a very low specificity.

The best use of these highly sensitive tests may be to rule out infection if they are both normal. Spangehl and associates used the combination of both ESR and CRP to achieve 100% specificity for excluding the diagnosis of infected THA.

ASPIRATION

Aspiration is accepted as an important tool in the workup for painful THA. However, its routine use before revision THA is more controversial. Roberts and associates reported sensitivity of 87% and specificity of 97% in their series of 78 patients who underwent revision THA. Levitsky and coworkers reported an even higher sensitivity of 92% and specificity of 97% when combined with an elevated ESR and a postoperative interval of less than 5 years. However, Barrack and Harris have been widely cited as finding, in a large series of 270 cases, a high false-positive rate of 13%, with only 6% positive predictive value, and have recommended against the use of aspiration routinely before every revision THA. Several authors have advised its use only after preliminary laboratory test results or nuclear imaging test results are abnormal, to confirm the diagnosis and identify the pathologic organism in question.

RADIOGRAPHY

Even in the absence of pain, postarthroplasty films are recommended at regular intervals to screen for early signs of wear or loosening. In the presence of pain, anteroposterior pelvic, anteroposterior hip, cross-table lateral, and frog-leg lateral views should be taken using the identical positioning as before. Serial examinations should follow the same technique, to compare subtle findings. A critical viewing includes assessment of component migration, component wear, component fracture, cement fracture, radiolucent lines, osteolysis, radiodense lines, sclerosis, pedestal formation, remodeling, and heterotopic ossification. Of note, the radiographic evaluation of cemented components versus uncemented components, acetabular components, and femoral components are all slightly different from each other.

When evaluating cemented femoral components for possible loosening, the recommendations of O’Neill and Harris have been widely accepted. They defined the following radiographic signs of loosening: migration of the component, cement, or component fracture or a continuous bone-cement radiolucent zone going all the way around the implant and that is wider than 2 mm at any point. In their retrospective series of over 60 hips, sensitivity was 89% and specificity was 100% (no false-positive findings), which suggests that this set of radiographic criteria can be reliably used to “rule in” a diagnosis of a loose cemented component. Miniaci and coworkers used a slight modification of these criteria in their series of 65 hips. They defined loosening as subsidence or migration greater than 1 cm, radiolucencies greater than 2 mm surrounding the entire component, or a progressive radiolucency surrounding 50% to 100% of the component. Using this set of radiographic criteria, sensitivity was 86% and specificity was 81% for femoral loosening.

The clinical significance of loosening at the prosthesis-cement interface, or debonding, varies greatly with the prosthesis design. Collarless, polished, and tapered femoral components have been shown in several series to be able to debond and subside slightly within the first several years to a stable position, with neither an increase in pain nor a decrease in survivorship. However, Mohler and associates demonstrated that if femoral components that are rough debond from the cement mantle, they can generate debris particles quickly and undergo rapid osteolysis.

Uncemented femoral components have been studied extensively by Engh and colleagues, who have identified major and minor radiographic signs of femoral component fixation and stability. The two major signs of osseointegration are the absence of reactive lines around the porous coated portion of the stem and the presence of bone “spot welds” bridging the endosteal bone and the porous coated portion of the stem. Minor signs of osseointegration are calcar atrophy, the absence of bead shedding, a stable distal stem, and the absence of a pedestal. Of note, the presence of a pedestal does not necessarily indicate loosening, unless associated with radiolucent lines about the smooth distal stem. A major sign of lack of osseointegration is the presence of extensive reactive, radiodense lines around the porous coated portion; however, the absence of endosteal spot welds is only considered a minor sign of lack of osseointegration.

The radiographic appearance of infection can be difficult to distinguish from that of aseptic loosening. However, some radiographic signs more specific to infection are endosteal scalloping, laminated periosteal new bone, and generalized osteopenia, especially if they progress rapidly during the first year. In their series, Lyons and colleagues reported endosteal scalloping had sensitivity of 47% and a specificity of 97% for infection and laminated periosteal new bone had a sensitivity of 25% and specificity of 92% for infection.

It is generally acknowledged that the acetabular side is more difficult to assess for radiographic loosening than the femoral side, because most of the acetabular interface is usually blocked from view by the component itself. Several series using different criteria for what defines a loose component have provided different results. O’Neill and Harris, when assessing the cemented acetabular components in their series, reported a sensitivity of only 37% using the radiographic criteria of either migration or a bone-cement radiolucent line that was continuous and 2 mm in width at some point. They also noted that by adjusting their criteria to be either migration or a radiolucent line that was continuous or 2 mm in width at some point the sensitivity did increase to 57%; however, owing to new false-positive findings that appeared, overall accuracy only rose from 63% to 68%. Miniaci and coworkers, using a modified Harris criteria that included any progressive radiolucency surrounding greater than 50% of the component, showed a sensitivity of 81% and specificity of 86% to predict acetabular loosening. Lyons and colleagues noted in their series that there was a 100% specificity for loosening (no false-positive findings) by using the radiographic criteria of migration, acetabular fracture or protrusio, or a bone-cement radiolucency that was 2 mm or more or progressively widening. Hodgkinson and coworkers reported that 94% of cemented sockets with a complete radiolucent line, regardless of width, were loose compared with 74% of cases with radiolucent lines in zones 1 and 2, 7% of cases with lines in zones 1 and 3, and only 5% of cases with lines in zone 1.

Southwell and associates described using oblique radiographs to bring more of the acetabular component interface into view. They reported that whereas a single anteroposterior view alone showed only 38% of the cup surface, adding 45-degree iliac and obturator oblique views showed 81% of the cup surface, and adding another 60-degree obturator oblique view showed 94% of the cup surface.

In regard to the evaluation of uncemented acetabular components for loosening, Berry suggested that screw fracture, or a complete radiolucent line, were of concern and that inferomedial lucency combined with superolateral sclerosis was suggestive of early tilting. Heekin and associates noted that bead shedding more than 2 years postoperatively correlates with migration and loosening. In their series of 52 total hip arthroplasties requiring reoperation, Udomkiat and colleagues noted that progression of radiolucent lines after 2 years and radiolucent lines greater than 1 mm that initially appeared after 2 years were both 100% predictive of loosening. More recently, Engh and coworkers noted from autopsy retrievals that acetabular ingrowth is unpredictable, is random, and does not follow a consistent pattern. They subsequently proposed five radiographic signs for detecting radiographic osseointegration of the porous coated acetabular component: absence of radiolucent lines, presence of a superolateral buttress, medial stress shielding, radial trabeculae, and presence of an inferomedial buttress. In their series of 119 total hip arthroplasties requiring revision surgery, Engh and coworkers reported that using the criteria of having three or more of these radiographic signs as criteria, the sensitivity for loosening was 90%, the specificity was 77%, and the positive predictive value was 97%.