The sample size was estimated on the basis of a precision of 4%, a long-lasting fixation of primary hip replacements using a metaphyseal fit modular of 95%, after 10 years of follow-up, an 80% confidence level, and losses of 10%. As a result, 54 patients were included in the study.

Between January 1998 and December 1998, a total of 114 total hip arthroplasties were performed in 105 patients at our institute. The mean age of patients was 78 years (75 – 86), and were implanted with 60 femoral components (six bilateral). Mean follow-up was 10,4 years (range, 10 – 11 years). Four patients (five hips) died before ten years from surgery due to unrelated cause. Two patients (one undergoing bilateral arthroplasty) were lost to follow-up. This left 48 patients whose data were retrospectively assessed and in whom 52 femoral components had been implanted. Of these 48 patients, 21 were men (43.75%) and 27 women (56.25%). 80% (n = 42) had been diagnosed with primary arthrosis, 8% (n = 4) with rheumatoid arthritis, and the the rest with (n = 6) idiopathic avascular necrosis.

All the patients received the same cementless modular stem. This stem formed part of the OMNIFLEX hip system (Howmedica Osteonics, Allendale, NJ), and was one of the early cementless designs. The modular femoral component system was designed to address the issue of proximodistal size mismatch and favor proximal stress transfer. Initially, the femoral component was made of Ti alloy with a double-wedge configuration and sintered proximal Ti bead pads, which were later replaced with a proximal circumferential arc-deposition with a 50 μm layer of hydroxyapatite (HA) coating (third-generation). The distal two-thirds of the component was tapered, to increase flexibility, with a grit-blasted coating, and the cylindrical polished CoCr tip was modular, to fit several distal canal diameters. A modular collar was available for placement after stem insertion, but in none of the present cases was this collar used. The intraoperative requirement for the use of the OMNIFLEX stem was that the test rasp could be stably fitted otherwise a cemented stem was used.

The acetabular component used was a semispherical Titanium Omnifit PSL (Howmedica Osteonics, Allendale, NJ) with a 10° polyethylene rim, to increase the superior-lateral cover and a 28 mm chrome-cobalt head.

In all cases, the surgical approach was the modified Hardinge transgluteal anterolateral, with the patient in a lateral position.¹⁴ A second-generation cephalosporin was administered half-an-hour preoperatively and during the first two postoperative days. On the second postoperative day, touch-toe weight-bearing walking with two crutches was allowed. Partial weight-bearing was encouraged at four weeks, gradually increasing to full weight-bearing at approximately 12 weeks postoperatively.

The patients were evaluated at three, six, and 12 months postoperatively and yearly thereafter, according to the Harris Hip Score.¹⁵ At their latest follow-up visit, all the patients underwent a detailed physical evaluation (BJT) and were checked for pain in the middle third of the thigh.

Radiographic follow-up was performed at the same time as the clinical follow-up and two of us (BJT, SZ) analyzed them at each follow-up. Preoperative assessment was based on anteroposterior and lateral views of the hip, including half the femur. Femoral morphology was determined by calculating the femoral canal index, defined as the ratio of the intracortical width of the femur, at a point 20 mm proximal to the tip of lesser trochanter and at the canal isthmus.¹⁰ Ratios lower than 3.0 indicated canals in the shape of a tube (type C femur), ratios of 3.0 – 4.7 indicated a normal canal (type B femur), and those greater than 4.7, a canal shaped like a champagne glass (type A femur).¹⁰

Using the radiographs obtained three months postoperatively, the proximal portion of the stem fill was measured at the central level of the lesser trochanter. This calculation of the metaphyseal filling was undertaken according to the modified criterion established by Dorr¹⁶ [Figure 1]. The position of the stem was recorded in relation to the anatomic axis of the femur (varus, valgus or neutral).

The fixation state of the femoral component was determined on an anteroposterior (AP) and lateral view of the hip obtained on the most recent follow-up visit. For this, we used the criteria established by Engh¹⁷ for femoral components with proximal porous coatings and classified the stems as stable with bone ingrowth (osseointegration), unstable with fibrous ingrowth. Analysis of the zones of bone ingrowth was undertaken according to the zones of Gruen,¹⁸ with Gruen zones 1, 2, 6, 7, 8, and 14 corresponding to the porous implant surface.

Bone atrophy in the proximal femur, due to stress shielding, was examined in the AP view of the hip obtained in the most recent follow-up, and graded according to the severity classification of Engh.¹⁹

Finally, we evaluated polyethylene (PE) wear using the technique described by Livermore.²⁰ The mean PE wear value was determined in the anteroposterior hip radiograph by calculating the difference between PE thickness in the most recent follow-up and the immediate postoperative, and dividing this figure by the number of years of follow-up. PE wear in the revised stems was correlated with the initial metaphyseal fill of the implant, to try to relate the reduced initial metaphyseal filling with the passage of debris particles due to wear. Similarly, we assessed whether the femurs showed more atrophy (osteopenia) on the anteroposterior view due to the shielding, which displayed greater PE wear.

Concerns about the use of cementless stems in elderly persons are essentially based on two facts. First, the high incidence of intraoperative fractures, and second, failure to achieve good bone ingrowth in the stem. The latter is the outcome of a poorer bone tissue quality, as this factor diminishes with age.^25,26 Accordingly, many surgeons consider the use of cementless stems in elderly persons inappropriate, on the grounds than an initial stem stability cannot be achieved.^27,28

No intraoperative fractures were recorded in our series. The essential prerequisite for the use of a cementless stem was that the test rasp could be stably fitted without the need for a cortical contact around the entire stem periphery. Four periprosthetic fractures, three type B1 and one type B2 according to the Vancouver classification, were recorded during the follow-up, and the stems were revised because the fracture could affect the stability of the implant. In some cases we used long modular stem revision (Restoration modular system, Stryker), and in other cases we used long stem revision (Restoration HA). Stem revision for oblique or transverse B1 fractures is now considered as a viable treatment modality, as this fracture configuration is difficult to control with single plating²⁹ [Figure 5].

No detrimental effects of fixation were observed in terms of pain and functionality. The incidence of pain, albeit low, recorded in the middle third of the thigh (3.8%) could be due to the diameter of the distal centralizer placed at the tip of the stem. The size selected for this centralizer was a diameter of at least 1 mm less than the last rasp used to assess the canal size.

Noble et al.¹⁰ established a radiographic classification scheme for femur canals, which was further developed by Dorr et al.³⁰ according to radiographic and histomorphometric criteria. Thus, following Noble's radiographic scheme and his premise that a single stem design would not be compatible with the different femoral canal types, we assessed the behavior of a single stem type, regardless of the morphology of the femoral canal.

In March 1997, the OMNIFLEX hip system was introduced in our Orthopedic Surgery Unit as the only uncemented femoral stem. Implants coated proximally with HA have provided good clinical outcomes, mostly in young patients.^12,31,32 However, only a few studies have assessed the performance of stems whose metaphyseal portion is coated with HA, in elderly patients. The present study has analyzed patients older than 75 years undergoing surgery in 1998 (52 stems implanted in 48 patients) after the initial learning curve. The use of stems had started several months prior to this date.

Our series of patients revealed a predominance of type B femoral canals. One stem was revised because of mechanical loosening. The initial metaphyseal fill was less than 70%, and its alignment was varus by 5°. Dorr¹⁶ and Martell³³ observed that metaphyseal fitted stems needed a fill exceeding 90% that required cortical contact. Our series of patients showed a mean metaphyseal fill of less than 90%, as we tried to avoid cortical contact over the entire periphery, to reduce the risk of intraoperative periprosthetic fracture. This could explain why HA coated stems need a lesser proximal fill.^34,35

Four of the femurs were type C, for which the use of cemented stem would normally be recommended.^36,37 Hence, obtaining sufficient stability for an optimal outcome of the implant of a cementless prosthesis, in this type of femur, is a particular challenge. Although the number of cases assessed here is too low to draw any valid conclusions, it would seem that the osteoconductive properties of HA might help stem ingrowth in type C femurs [Figure 6].

The results obtained with the modular OMNIFLEX system have been reviewed by several authors in younger patients, but to our knowledge, no series describing the outcome of this stem have been published in elderly patients. Capello, Kitamura, and Ito^38,39,40 reported poor results using first generation stems, which featured a proximally non-circumferential porous tip. However, Takahashi³⁵ obtained good results with the use of second generation stems, with arc deposition of pure titanium on the surface of the proximal circumference, and with third generation models whose circumferential porous tip was coated with HA. No revision stem was required, and 97% of all stems showed bone ingrowth fixation. We report one revised stem (1.9%) due to aseptic loosening. Thus there are clear differences among the results obtained with the first, second, and third generation stems, with the latter showing the best outcomes. In effect, a HA-coated metaphyseal portion promotes bone ingrowth.

In our series, 22% of the stems showed a valgus or varus alignment, with no functional repercussions. This could be explained by the size of the distal centralizer used, which was at least 1 mm smaller in diameter than the size of the last rasp of the centralizer. Our findings indicated that despite mal-alignment, there was adequate metaphyseal fill and the HA coating promoted bone ingrowth, with no clinical significance [Figure 4]. On the other hand, the present stem revised because of subsidence was attributed to a technical error in establishing a metaphyseal fill, besides its varus position. In effect, Capello³⁸ highlighted the technical skill needed for the use of this hip system.

Bone atrophy of the proximal end of the femur following total hip arthroplasty is widely established.⁴¹ We recorded shielding-related bone atrophy in 75% of the hips. This adaptive process was observed during the first few years of follow-up, yet there was no progression thereafter. Cancellous condensation indicating osseointegration in stems, stable by bone ingrowth, mainly appeared in Gruen zones 2 and 6, which explained the high proportion of proximal bone atrophy. This behavior was typical of stems with a greater surface area for bone ingrowth, such as, extensively porous stems. Thus, we could infer that OMNIFLEX stems became more distally integrated in the host tissues despite their porous circumferential surface coated with HA, which circumscribed the metaphyseal zone, probably due mainly to the grit-blasted finish of the middle third of the stem. Takahashi³⁵ reported proximal bone atrophy in 65% of the second generation and 68% of the third generation stems, in patients of a mean age of 64 years. The mean age of our patients was above 75 years. Younger, more active patients with better bone quality develop less atrophy due to stress shielding. The three stems showing stable fibrous ingrowth exhibited grade I atrophy, which could be attributed to the fibrous fixation. We noted atrophy below the lesser trochanter without diaphysis involvement (third grade) in 16% of the cases. This incidence was lower than that reported by others.^41,42 The OMNIFLEX stem was shaped like a double wedge and was less stiff, explaining its behavior. Notwithstanding, according to the other authors,⁴³ we were aware of the radiographic limitations of assessing bone mineral density loss, and concur with Glassman⁴⁴ who stresses on the importance of the use of DEXA (dual energy X-ray absorptiometry) in this field.

Polyethylene wear was not greater in the revised stem because of mechanical loosening. As shown in other studies, neither was this PE wear greater in femurs showing atrophy due to stress shielding,⁴⁵ precluding its relation with osteolysis induced by debris.