Obtaining a satisfying long-term survival of THAs in patients younger than 40 years remains a challenge. Young patients must function longer with their THA than the typical patient who has a THA, and they also engage in a higher level of activity, which is associated with higher revision rates [19, 27]. Therefore, this population is more dependent on durable implants with excellent long-term survival. Although stem survival is acceptable in most studies, in general, cup survival is the weakest link in patients younger than 40 years [5, 8, 9, 15, 18, 20, 25]. The difference in reported survival rates between the cup and stem varies from 1% (97% [stem] versus 96% [cup] [18]) to 11% (98.3% [stem] versus 87.6% [cup] [20]). Despite attempts to improve cup designs and using new materials in THA, the acetabular component still shows lower survival rates than femoral implants.

One popular option is to implant uncemented acetabular cups in young patients as part of a total uncemented THA or hybrid THA (uncemented cup, cemented stem). Although cement in young patients commonly is not used [1, 24, 35], we always have implanted cemented cups in patients of all ages, but with one substantial modification: in all patients with substantial acetabular bone stock deficiencies, we have reconstructed this bone stock loss using impaction bone grafting with a cemented cup. Secondary osteoarthritis resulting from underlying diseases in these young patients often is seen with associated loss of acetabular bone stock (for example, in developmental dysplasia of the hips and juvenile rheumatoid arthritis).

With this approach using cemented cups in young patients for many years, we asked whether there were any differences between cemented cups in young patients (younger than 40 years) with and without reconstruction with impaction grafting concerning (1) clinical scores, (2) revisions, (3) complications, (4) radiographic appearances, (5) polyethylene wear, and (6) survival.

We categorized acetabular defects in accordance with the classification system of the American Academy of Orthopaedic Surgeons [10]. Eighty-six hips (49%) had an acetabular deficiency. Type I segmental deficiencies occurred in 16 hips, Type II cavitary defects in 39 hips, and Type III combined deficiencies in 29 hips. One patient (two hips) had ankylosis of the hips, a Type V deficiency. Using impaction grafting, we reconstructed all deficiencies, including mild cavitary defects; however most were larger defects.

Differences between the two groups (with and without impaction grafting) were analyzed regarding diagnosis and gender (chi square test, both p = 0.001). In the group with an acetabular reconstruction, a larger proportion was female and was diagnosed with developmental dysplasia of the hips compared with the group without reconstruction. There were no differences regarding age at surgery, side, bilateral THAs, followup, type of cup used, cup inner diameter, and body mass index between the two groups.

Two-thirds of the operations (67%) were performed by or under the supervision of two senior faculty orthopaedic surgeons (BWS, JWMG). A posterolateral approach without trochanteric osteotomy was used in all hips, with the exception of two. Intraoperatively, in one patient, a preplanned Sugioka procedure was converted to a THA; however, a trochanteric osteotomy already had been performed. In the other patient, a trochanteric osteotomy was performed in a technically demanding hip with a short femoral neck. In one patient, an additional anterior approach was needed because of ankylosis of the hip. All acetabular deficiencies were reconstructed (with the exception of one case) with impaction grafting using autografts and/or allografts in 84 hips (48%); this technique has been described in detail [28–30]. Segmental bone defects first were reconstructed with wire meshes before the morselized bone graft was impacted and a conventional full polyethylene cup was cemented. In one patient, we reconstructed a lateral rim deficiency without impaction grafting using a solid autograft fixed with two screws. In one of the ankylosed hips (Type V deficiency), we did not use impaction grafting. We used allografts only with impaction grafting in four hips (4.8%), autografts only in 72 hips (85.7%), and combined allografts and autografts in eight hips (9.5%). Allografts were used when the original femoral head was not large enough to reconstruct the defect or in cases with pathologic femoral heads (for example, avascular necrosis of the femoral head). In three cases, instead of a solitary metal mesh, a solid fragment was used in combination with impaction grafting. In two of these cases, a minor segmental defect in the medial wall was closed using a cortical-trabecular fragment of a femoral head. A wire mesh was placed medial on top of the fragment and the acetabulum was reconstructed with impaction grafting. In the third case, a cortical head fragment was used to support the anterior rim together with a rim mesh in a reconstruction. The number of femoral heads used as grafts varied from one to four. In 40 hips (48%), metal wire mesh was used for acetabular reconstruction with impaction grafting (10 medial wall meshes, 39 rim meshes). In nine early cases, we placed a mesh on top of the bone graft just before cementation, but this mesh was not part of a segmental defect reconstruction. However, after we realized this mesh did not add any stability to the reconstruction and there were no signs of damaging of the graft or graft healing by direct contact with cement, we abandoned the use of a mesh for this purpose.

We used 79 (45%) Exeter™ Contemporary™ cups with an inner diameter of 28 mm (n = 75) and 22.225 mm (n = 5) (Stryker Howmedica, Newbury, UK), 71 (41%) Charnley^® Elite™ cups with an inner diameter of 22.225 mm (n = 6) or 28 mm (n = 65) (DePuy, Leeds, UK), and 25 (14%) Müller/AlloPro cups with an inner diameter of 32 mm (n = 19), 28 mm (n = 2), or 22.225 mm (n = 4) (Sulzer, Winterthur, Switzerland). For the femoral component, we used an Exeter™ stem in 111 cases, a Charnley^® Elite™ stem in 48 cases, and a Müller stem in 16 cases. All femoral heads used were made of a cobalt-chrome alloy; no ceramic implants were used.

We cemented acetabular components with a third-generation cementing technique. In the directly cemented cups, after reaming, multiple small drill holes were made with a 2.6-mm drill. After using pulse lavage, vacuum-mixed cement was injected directly from the cement gun and the cement was pressurized by a seal. In cases of reconstruction with bone grafts, we reamed the acetabulum, made multiple drill holes in sclerotic areas, and irrigated the acetabulum. Next the bone graft was impacted. Again, vacuum-mixed cement was injected and pressurized and the cup was inserted. Before 1989, we used Palacos^® bone cement (Merck, Darmstadt, Germany); however, since 1989, we have used Surgical Simplex^® (Stryker Howmedica). In 165 cases (94%), cement loaded with antibiotics was used. All patients received antibiotic prophylaxis consisting of 2 g cefazolin intravenously just before surgery. Other precautionary measures to prevent infections were use of an operating theater with laminar airflow and use of two pairs of sterile gloves.

Postoperatively, all patients received thrombosis prophylaxis with low-molecular-weight heparin for 6 weeks, or before 1999, with acenocoumarol (the individual dosage regimens regulated with regular coagulation tests) for 3 months. To prevent heterotopic ossification, we used nonsteroidal antiinflammatory drugs (NSAIDs) for 7 days. In six patients in whom NSAIDs were contraindicated, we administered one dose (7 Gy) of radiotherapy 1 day postoperatively.

Patients without acetabular reconstruction were mobilized under supervision of a physiotherapist after 1 or 2 days. Full weightbearing was increased in 2 to 6 weeks with the aid of one or two crutches. The patients who underwent impaction grafting were mobilized according to a modified protocol; in the first 6 weeks, only 10% weightbearing was allowed and then 6 to 12 weeks of 50% weightbearing using two crutches was allowed. After 12 weeks, full weightbearing mobilization was allowed. Thirty-one hips had such an extensive reconstruction of major defects that several weeks of bed rest were maintained ranging from 1 to 6 weeks. We used this modified mobilization protocol to ensure graft incorporation before full weightbearing.

Routine followups were scheduled at 6 weeks; 3, 6, and 12 months; and yearly or biannually thereafter. At our outpatient clinic, student researchers not participating in the treatment performed clinical analysis using the Harris hip score [17], the Oxford Hip Questionnaire Score (since 1998) [11], and visual analog scales for pain during rest and physical activity on a scale from 0 (no pain) to 100 (unbearable pain). We report the clinical scores of all patients excluding the 21 patients whose hips were revised during followup.

All anteroposterior pelvis and lateral radiographs of all hips were analyzed on a consensus basis by two of the authors (DCJDK, BWS). Radiographic evaluation included assessment of cup position, loosening of the acetabular component, polyethylene wear, presence of osteolysis, structural quality of the bone graft, application and position of the meshes, migration, heterotopic ossification, and fracture of the cement, mesh, or prosthesis. Radiolucent lines and osteolysis were recorded according to the three acetabular zones as described by DeLee and Charnley [12]. Radiographic loosening was defined as 2 mm or greater demarcation in two or three zones around the acetabular component, progressive demarcation, 3 mm or greater component migration, 5° or greater component tilting, and/or cement or prosthesis fracture. We determined cup migration (> 3-mm shift in any direction or > 5° tilting) in relation to the interteardrop line instead of the Kohler line [16]. Position of the cup of 45° ± 10° was considered normal [26]. We calculated polyethylene wear using the method of Dorr and Wan [13]. All measurements were corrected for magnification. Heterotopic ossification was classified according to the system of Brooker et al. [6]. Graft incorporation was defined as the presence of the crossing of trabecular bone on the bone-graft interface on the radiographs. Clinical failure was defined as the need for revision of the acetabular component for any reason.

We calculated Kaplan-Meier curves to study the survival (time to revision). The end points were (1) cup revision for any reason, (2) cup revision for any reason excluding infections, (3) cup revision for aseptic loosening, and (4) radiographic signs of cup loosening. With an average followup of 8.1 years, 30% of all patients had a followup longer than 10 years. The log-rank test was used to test the differences in survival between cups with and without impaction grafting. Differences in outcomes between the groups were determined with the Student’s t-test (continuous variables after checking for normal distribution) or chi square test (nominal variables).

There were no differences between the cups with and without acetabular reconstruction concerning the occurrence of cup migration, radiographic loosening, or the presence of osteolysis, cysts, and abnormal cup position (Table 5). Cups with impaction grafting had fewer radiolucent lines (p = 0.02) and fewer lines in Zone I (p = 0.001) (Table 5). All lines, except two, were on the bone-cement interface. In 28 (48%) of the 58 cups with radiolucent lines, the lines were progressive. Of the 175 hips, 160 were radiographically stable (Fig. 1). Fifteen cups were difficult to evaluate because of overlap of the metal mesh (11 Zone I; four Zones I + II). We observed graft osteolysis in only one patient with impaction grafting; all other grafts were fully incorporated. The hip revised because of traumatic loosening had a fracture in Zone II of the acetabulum; no other fractures were seen. Fifteen (8.6%) cups were radiographically loose, three had cup migration (after 1.8, 9.8, and 11.2 years postoperatively), and 12 had evident radiolucent lines in all zones and/or severe osteolysis; 12 of these cups were revised (Table 3).

Table 5 Radiographic findings of all cups*

Radiographic finding	All	With bone impaction grafting	Without bone impaction grafting	p Value (where appropriate)
Radiographic loosening	15	5	10	0.608
Cup migration	3	1	2	0.234
Radiolucent lines	58	18	40	0.02
Zone I	18	3	15	0.001
Zone II	2	0	2
Zone III	17	9	8
Zones I + II	4	2	2
Zones II + III	5	2	3
Zones I + III	5	1	4
Zones I + II + III	7	1	6
Osteolysis	11	5	6	0.861
Zone I	7	2	5
Zone II	1	1	0
Zone III	3	2	0
Cysts	1	0	0	0.033
Zone I	1	0	0
Zone II	0	0	0
Zone III	0	0	0
Cup position
Neutral position (35°–55°)	160	77	83
Abnormal position	15	7	8	0.914
Vertical (> 55°)	12	6	6
Horizontal (< 35°)	3	1	2
Polyethylene wear
Mean nonrevised cups (mm/year)	0.080	0.076	0.084	0.539
Mean revised cups (mm/year)	0.214	0.182	0.230	0.525

* Total cups (n = 175); Cups with (n = 84) or without (n = 91) reconstruction with bone impaction grafting.

Fig. 1A–C The radiographs illustrate reconstruction of the acetabuli in a 34-year-old woman with bilateral DDH (Crowe Grade 3). (A) A preoperative anteroposterior radiograph shows the acetabuli. (B) An anteroposterior radiograph taken immediately postoperatively shows the THAs with the acetabuli reconstructed with impaction grafting. (C) An anteroposterior radiograph taken 12 years postoperatively shows the THAs remain radiographically stable, but Brooker Classes III (left) and I (right) heterotopic ossifications are visible.

---

There was no difference in polyethylene wear rates between the cups with and without impaction grafting (p = 0.539 in 154 unrevised cups and p = 0.525 in the 21 revised cups) (Table 5). When looking at all cups (with and without acetabular reconstruction), the revised and radiographically loose cups had more wear compared with the cups that were not revised (both p < 0.0001). Patients with an abnormal position of the cup had similar (p = 0.196) polyethylene wear rates to those who had a normal position. Analysis of polyethylene wear rates of cups with different inner diameters showed no differences (independent t test, 22 versus 28 mm: p = 0.135, 22 versus 32 mm: p = 0.484, 28 versus 32 mm: p = 0.620).

There were no differences in survival after 10 years between the groups with and without bone impaction grafting (Table 6). The midterm survival rates of all cemented polyethylene cups varied from 85% to 92% at 10 years with four end points (Table 6; Figs. 2, 3). Cup survival with an end point of radiographic loosening was 89% (95% confidence interval, 83%–95%).

Table 6 The 10-year survival rates*

End point	All cups	Without bone impaction grafting	With bone impaction grafting	Log-rank p value
Revision for any reason	85% (78%–92%)	79% (68%–90%)	91% (82%–99%)	0.21
Revision for any reason excluding infections	91% (85%–97%)	87% (78%–99%)	94% (87%–100%)	0.56
Revision for aseptic loosening	92% (87%–98%)	90% (81%–99%)	95% (89%–100%)	0.73

* Kaplan-Meier estimates; 95% confidence interval in parentheses.

Fig. 2A–B Kaplan-Meier survival curves with 95% confidence intervals (broken lines) of all cups with end points of (A) revision for any reason and (B) revision for aseptic loosening are shown. The vertical bars indicate the censored data points.

---

Fig. 3A–B Kaplan-Meier survival curves of cups without impaction grafting (thick broken line, 95% confidence intervals in thin broken lines) and cups with impaction grafting (thick solid line, 95% confidence intervals in thin solid lines) with end points of (A) revision for any reason and (B) revision for aseptic loosening are shown.

---

The use of cemented THA in young patients is not very popular and most surgeons will use uncemented or will resurface hips in these patients. However, we have continued to use only cemented implants in THA even in young patients. In our view, the real challenge in THA in these young patients is to manage the commonly seen acetabular deficiencies. In cases of acetabular defects, we reconstruct these deficiencies with impaction bone grafting. We questioned whether there was a difference in clinical outcome, revisions, complications, radiographic appearances, polyethylene wear, and survival between the cups implanted with an acetabular reconstruction with impaction grafting and those implanted with standard cementing techniques.

Our study has several limitations: short followup, lack of assessment of activity levels, clinical interobserver variability, heterogeneous group, no comparison with other reconstruction techniques, and different types of implants used. With no patients lost to followup, our followup is representative and reliable for the midterm results [22], and long-term followup (> 15 years) was not available at the time of this review. Our results can be biased by an important factor we did not evaluate: the level of activity. Theoretically, with restoration of the affected hip(s) into well-functioning artificial joints, most patients will increase their level of activity. However, young patients undergoing THA with acetabular deficiencies and therefore more complex reconstructions could still have a lower level of activity after surgery relative to primary cemented cups. However, the average wear of the cups with impaction grafting was the same as the cups without impaction grafting (both 0.08 mm/year). Provided that activity is a major cause of polyethylene wear, this might imply the level of activity is similar in these two groups. Several studies suggest the revision and polyethylene wear rates are correlated to level of activity [2, 19, 27, 31, 38]. Additional research on level of activity and impaction grafting in young patients is necessary to confirm this hypothesis. The clinical questionnaires were obtained by student researchers who did not participate in the treatment. Multiple researchers were involved in the data collection and interobserver variability has not been tested; however, all researchers were trained and supervised to obtain these questionnaires correctly.

Although revision rates in both groups were comparable, the time to revision was longer in the cups reconstructed with bone impaction grafting. We have no clear explanation for this observation; possibly the cement-bone interface was better in cups with bone impaction grafting with better interdigitation of the cement into the bone [36]. This also may explain the lower incidence of radiolucent lines in the cups reconstructed with bone impaction grafting. The number of revisions for septic loosening was relatively high during this midterm followup study (4.6%). Only one septic loosening likely was related to the surgery; we considered all other infections acute hematogenous infections of previously well-functioning prostheses. The use of corticosteroids and newer rheumatic disease-modifying drugs, which were used in most of the infection cases, can explain this higher risk of infection [3, 4]. Sochart and Porter [33] had only two infections in their study, but both were in patients with rheumatoid arthritis. Still, our revision rate for septic loosening of 4.6% is relatively high in contrast to other studies, such as that of Joshi et al. [18], with an infection rate of 1.3%. Remarkably, many septic loosenings occurred late (> 2 years postoperatively).

The cups reconstructed with impaction grafting showed fewer complications by having fewer dislocations than the cups implanted by standard techniques. This might be attributed to the different mobilization protocol for the patients who received cups with impaction grafting. Immobilization is associated with lower dislocation rates [21]. The overall dislocation rate in our study was 11.4%, which is relatively high. However, subluxation rates in young patients having THA have been reported to be as much as 18.2% [14]. The overall complication rate of 17% (30 postoperative complications) is also relatively high. Joshi et al. [18] reported a complication rate of 11.5% in cemented hips and Duffy et al. [15] reported a complication rate of 12% during the perioperative period.

As expected, revised and radiographically loose cups showed more polyethylene wear. This is consistent with previous reports showing wear particles are associated with osteolysis in THA [27, 33, 37]. The average wear rate of the cups of 0.08 mm/year is within the normal limits, keeping in mind that wear in younger patients can be 33% to 40% higher than wear in older patients [27]. In a large study of 226 hips in patients younger than 40 years with a Charnley^® prosthesis, Sochart and Porter [33] reported an average wear rate of 0.08 to 0.10 mm/year in the nonrevised cups, which is comparable to our results. Wan et al. [37] found a correlation between inclination of the cup and higher/lower wear rates. However, we did not observe higher wear with abnormal position or inner cup diameter.

A remarkable finding of our study was the survival of cups with acetabular reconstructions with impaction grafting was at least comparable to the survival of standard cemented cups, especially considering the more difficult hips of our study population needed reconstruction with impaction grafting. Our data on the cemented cups with impaction grafting showed similar survival, where rather lower survival rates would be expected. The outcome of these cups reconstructed with impaction grafting even fulfilled the NICE criteria (a survival of > 90% after 10 years) [23], with a survival rate of 91% at 10 years with an end point of revision for any reason. The survival rates of the cemented cups in our study are comparable to those reported for cemented cups [8, 18, 34].

Although cemented cups are not commonly used in young patients, our data suggest cemented conventional polyethylene cups are still a good option in THA in young patients. Even reconstruction of (severe) acetabular deficiencies with impaction grafting and a cemented conventional polyethylene cup produced very acceptable survival rates, comparable to the rates of cemented cups implanted in acetabuli without deficiencies with standard cementing techniques.