Fracture risks are the most common argument against the use of Ceramic on ceramic (CoC) hip implants. Question: is ceramic material at risk in case of severe local trauma? Over a long period, we tried to identify patients with a CoC prosthesis (Ceraver Osteal°)who did sustain a trauma. This was conducted in three different institutions. Eleven patients were found: 9 males and 2 females aged 17 to 70 years at time of index surgery. Accident occurred 6 months to 15 years after index: one car accident, five motorcycle accident, five significant trauma after a fall, including one ski board accident. Consequences of these trauma were: six fractures of the acetabulum with socket loosening in 4 that needed revision, two femoral shaft fracture, one orifed and one stem exchanged, one traumatic hip dislocation associated to loosening of the socket revised at 10 years, and one traumatic loosening of the socket. Ten had no consequence on ceramic integrity. One experienced a fracture of the patella from a dashboard trauma, a liner shipping was discovered during socket revision 2 years later. This is the only case of possible relation between trauma and ceramic fracture. In a more recent longitudinal study on 1856 CoC prosthesis performed from 2010 to 2021, 29 severe traumas were identified with no consequence on Ceramic material. From this limited case study, it can be assumed that Pure Alumina Ceramic well designed and manufactured, will not break after a significant trauma.
Bone stock restoration of acetabular bone defects using impaction bone grafting (IBG) in total hip arthroplasty may facilitate future re-revision in the event of failure of the reconstruction. We hypothesized that the acetabular bone defect during re-revision surgery after IBG was smaller than during the previous revision surgery. The clinical and radiological results of re-revisions with repeated use of IBG were also analyzed. In a series of 382 acetabular revisions using IBG and a cemented component, 45 hips (45 patients) that had failed due to aseptic loosening were re-revised between 1992 and 2016. Acetabular bone defects graded according to Paprosky during the first and the re-revision surgery were compared. Clinical and radiological findings were analyzed over time. Survival analysis was performed using a competing risk analysis.Aims
Methods
Pelvic tilt can vary over time due to aging and the possible appearance of sagittal spine disorders. Cup position in total hip arthroplasty (THA) can be influenced due to these changes. We assessed the evolution of pelvic tilt and cup position after THA and the possible appearance of complications for a minimum follow-up of ten years. 343 patients received a cementless THA between 2006 and 2009. All were diagnosed with primary osteoarthritis and their mean age was 63.3 years (range, 56 to 80). 168 were women and 175 men. 250 had no significant lumbar pathology, 76 had significant lumbar pathology and 16 had lumbar fusion. Radiological analysis included sacro-femoral-pubic (SFP), acetabular abduction (AA) and anteversion cup (AV) angles. Measurements were done pre-operatively and at 6 weeks, and at five and ten years post-operatively. Three measurements were recorded and the mean obtained at all intervals. All radiographs were evaluated by the same author, who was not involved in the surgery.Introduction
Materials and methods
Biological repair of acetabular bone defects after impaction bone grafting (IBG) in total hip arthroplasty could facilitate future re-revisions in case of failure of the reconstruction again using the same technique. Few studies have analysed the outcome of these acetabular re-revisions. We analysed 34 consecutive acetabular re-revisions that repeated IBG and a cemented cup in a cohort of 330 acetabular IBG revisions. Fresh-frozen femoral head allografts were morselized manually. All data were prospectively collected. Kaplan-Meier survivorship analysis was performed. The mean follow-up after re-revision was 7.2 years (2–17). Intraoperative bone defect had lessened after the first failed revision. At the first revision there were 14 hips with Paprosky 3A and 20 with Paprosky type 3B. At the re-revision there were 5 hips with Paproky 2B, 21 with Paprosky type 3A and 8 with type 3B. Lateral mesh was used in 19 hips.Introduction
Patients and Methods
Aseptic loosening is rare with most cementless tapered stems in primary total hip arthroplasty (THA), however different factors can modify results. We ask if the shape and technique of three current different femoral components affects the clinical and radiological outcome after a minimum follow-up of ten years. 889 cementless tapered stems implanted from 1999 to 2007 were prospectively followed. Group 1 (273 hips) shared a conical shape and a porous-coated surface, group 2 (286 hips) a conical splined shape and group 3 (330 hips) a rectangular stem. Clinical outcome and anteroposterior and sagittal radiographic analysis were compared. Femoral type, stem position, femoral canal filling at three levels and the possible appearance of loosening and bone remodelling changes were assessed.Background
Methods
Impaction bone grafting (IBG) is a reliable technique for acetabular revision surgery with large segmental defects. However, bone graft resorption and cup migration are some of the limitations of this tecnique. We assess frequency and outcome of these complications in a large acetabular IBG series. We analysed 330 consecutive hips that received acetabular IBG and a cemented cup in revision surgery with large bone defects (Paprosky types 3A and 3B). Fresh-frozen femoral head allograft was morselized manually. The mean follow-up was 17 years (3–26). All data were prospectively collected. Kaplan-Meier survivorship analysis was performed. Changes in different paremeters regarding cup position were assessed pre- and postoperatively and at the follow- up controls. Only variations greater than 5º and 3 mm were considered.Introduction
Patients and Methods
Acetabular fractures management is controversial since, despite a good anatomical reduction, clinical outcome is not satisfactory very often and the probability of a total hip arthroplasty (THA) is high. Surgical treatment include long operating times, large approach, blood loss, neural and muscle damage, and a high risk of failure and secondary osteoarthritis related to bone necrosis, cartilage damage, and bone loss. We hypothesized that the acetabular fracture management affected the clinical and radiological outcome of THA after posttraumatic arthritis. We compared 49 patients (49 hips) initially treated conservatively followed some months later by THA in conjunction with acetabular reconstruction (group 1); and 29 patients (29 hips) who had undergone THA after a failed osteosynthesis (group 2). There were more associated fractures according to Letournel in group 2. The mean age was 59.3±15.8 years for group 1 and 52.9±15.2 years for group 2. The mean delay between fracture and THA was 75.4±5 months for group 1 and 59.4±5 for group 2. The mean follow-up was 11.7 in group 1 and 10.2 in group 2. Preoperative bone defect was similar. We used bone autograft in 13 hips (26.5%) in group 1 and four (13.6%) in group 2. We used acetabular reconstruction plates in 2 hips with a pelvic discontinuity in group 1. Complications, clinical outcome according to Harris Hip Score, and radiological reconstruction were compared. Two-way ANOVA with repeated measures were used for comparison.Introduction
Materials and Methods
Different pathologies, deformities, bone defects, previous surgeries and polyethylene wear limit the survival of total hip arthroplasty (THA) in young patients. We compare preoperative status and outcome in 171 young and very young patients (207 hips) who underwent a ceramic-on-ceramic THA. Sixty-three (77 hips) were less than 30 years old (group 1) and 108 (130) were between 30 and 40 years old. Mean follow-up was 11.3 years. Two-way ANOVA with repeated measures were used to analyse clinical and radiological changes. Juvenile rheumatoid arthritis (JRA) was the most frequent diagnosis in group 1 and avascular necrosis in group 2. Charnley class type C, low activivity, previous surgery and osteoporotic bone were more frequent in group 1 patients. There were 2 cup revisions for aseptic loosening in group 1 and 4 in group 2. Survivorship analysis at 15 years was 96.7% (95% IC 92.2 a 100) for group 1 and 96.1% (95% IC 92.2 to 100) for group 2 (p=0.749). Despite the worse preoperative status in group 1 patients, clinical outcome was similar in both groups. Severe dysplasia had the worst clinical result and avascular necrosis the best. Patients with Charnley class C and JRA showed the most improvement. Radiographic reconstruction of the abduction angle was worse in group 1 (p=0.02). No osteolysis or complications derived from ceramic use were found. We conclude that despite the worse preoperative status in group 1 patients, clinical and radiographic results were good in both groups of patients who received a ceramic-on-ceramic THA.
Alumina-on-alumina in total hip replacement has been used for avoiding osteolysis and loosening. Published series report no ceramic wear and low rates for fractures and noises, but report poor results because of acetabular fixation failure. From 1999 to 2005, we used the “first generation” of a cementless cup, tri-radius relatively-smoothed HA coated (group 1), and from 2006 we have used a “second-generation” of this same cementless cup design with a macrotextured surface (group 2). We compare the perioperative conditions of two groups of patients using these two different cups and the clinical and radiological results. We analysed 679 (612 patients) consecutive and non-selected primary cementless alumina-on-alumina prostheses. There were 342 hips in group 1 and 337 in group 2. The stem used for all patients in this series was the same and fitted with an Al2O3 liner and femoral head. The use of screws were according to the intraoperative stability of the cup (pull-out test). Patients’ mean age was 48.7+13.6 years and the average follow-up until revision or the last evaluation was 11.7 years for group 1 and 5.4 years for group 2.Introduction
Material and Methods
Recent proposals have been introduced to modify stem design and/or femoral fixation in total hip replacement (THR). New designs need to consider previous design features and their results. The aim of this study has been to evaluate the clinical and radiological results of six different designs of tapered uncemented stems implanted in our Institution. 1918 uncemented hips were prospectively assessed from 1999 to 2011 (minimum follow-up of five years for the unrevised hips). All hips had a 28 or 32 mm femoral head and metal-on-polyethylene or alumina-on-alumina bearing surface. Six uncemented femoral designs that shared a femoral tapered stem incorporating a coating surface were included in the study. The different design features included the type of coating, metaphyseal filling, and sectional shape.Background and aim
Methods
The use of screws is frequent for additional fixation, however, since some disadvantages have been reported a cup press-fit is desirable, although this can not always be obtained. Cup primary intraoperative fixation in uncemented total hip replacement (THR) depends on sex, acetabular shape, and surgical technique. We analyzed different factors related to primary bone fixation of five different designs in patients only diagnosed with osteoarthritis, excluding severe congenital hip disease and inflammatory arthritis, and their clinical and radiological outcome. 791 hips operated in our Institution between 2002 and 2012 were included for the analysis. All cases were operated with the same press-fit technique, and screws were used according to the pull-out test. Two screws were used if there was any movement after the mentioned manoeuvres. Acetabular and femoral radiological shapes were classified according to Dorr et al. We analyzed radiological postoperative cup position for acetabular abduction angle, the horizontal distance and the vertical distance. Cup anteversion was evaluated according to Widmer and the hip rotation centre according to Ranawat.Introduction
Materials y Methods
Although cemented fixation provides excellent results in primary total hip replacement (THR), particularly in patients older than 75 years, uncemented implants are most commonly used nowadays. We compare the rate of complications, clinical and radiological results of three different designs over 75-years-old patients. 433 hips implanted in patients over 75 years old were identified from our Local Joint Registry. Group A consisted of 139 tapered cemented hips, group B of 140 tapered grit-blasted uncemented hips and group C of 154 tapered porous-coated uncemented hips. A 28 mm femoral head size on polyethylene was used in all cases. The mean age was greater in group A and the physical activity level according to Devane was lower in this group (p<0.001 for both variables). Primary osteoarthritis was the most frequent diagnoses in all groups. The radiological acetabular shape was similar according to Dorr, however, an osteopenic-cylindrical femur was most frequently observed in group A (p<0.001). The pre- and post-operative clinical results were evaluated according to the Merle-D'Aubigne and Postel scale. Radiological cup position was assessed, including hip rotation centre distance according to Ranawat and cup anteversion according to Widmer. We also evaluated the lever arm and height of the greater trochanter distances and the stem position. Kaplan-Meier analysis was done for revision for any cause and loosening. The hip rotation centre distance was greater and the height of the greater trochanter was lower in group B (p=0.003, p<0.001, respectively). The lever arm distance was lower in group C (p<0.001). A varus stem position was more frequently observed in group B (p<0.001). There were no intra- or post-operative fractures in group A, although there were five intra-operative fractures in the other groups plus two post-operative fractures in group B and four in group C. The rate of dislocation was similar among groups and was the most frequent cause for revision surgery (8 hips for the whole series). The mean post-operative clinical score improved in all groups. The overall survival rate for revision for any cause at 120 months was 88.4% (95% CI 78.8–98), being 97.8% (95% CI 95.2–100) for group A, 81.8% (95% CI 64.8–98.8) for group B and 95.3% (95% CI 91.1–99.6) for group C (log Rank: 0.416). Five hips were revised for loosening. The overall survival rate for loosening at 120 months was 91.9% (95% CI 81.7–100), being 99.2%(95% CI 97.6–100) for group A, 85.5 (95% CI 69.9 −100) for group B and 100% for group C (Log Rank 0.093).Materials and Methods
Results
Uncemented press-fit cups provide bone fixation in primary total hip replacement (THR). However, sometimes screws are needed to achieve primary stability of the socket. We analyzed biomechanical factors related to press-fit in seven cup designs and assessed whether screw use provides similar loosening rates to those of the press-fit technique. From a series of 1,350 primary uncemented THRs using seven different press-fit cup designs (a dome loading hemispheric cup and bi- or tri- radius cups), we only analyzed the 889 diagnosed of primary osteoarthritis. All cases were operated by the same surgical team. The use of screws was decided intraoperatively based on cup stability according to the pull-out test. There were 399 female and 490 male patients with a mean age of 65 years old. The mean follow-up was 8.6 years (5–13 years). The reconstruction of the hip rotation center was evaluated according to Ranawat.Introduction
Materials y Methods
Different bearing surfaces, including alumina-on-alumina have been used to avoid osteolysis. We ask if the use of modern ceramics matched with metal-backed sockets improves clinical results avoiding aseptic loosening, osteolysis and late dislocations. We prospectively analysed 315 Cerafit cups in two different generations: First generation implanted between 1999 and 2005 (124 cups) and second generation (191 cups) all implanted with a 5-year minimum follow-up, and associated with a Multicone-HAP stem using alumina-on-alumina bearings. Patients' mean age was 48.7+13.6 years. There were three dislocations within the first postoperative month and one early recurrent dislocation requiring reoperation using a double mobility cup. There were four intraoperative fractures, two in rheumatoid juvenile arthritis cases solved with a cerclage, and one postoperative fracture solved using a femoral long-stem. No fractures were seen in the alumina components in this series. Also, there have been no infections. There were 4 patients reporting non-reproducible squeaking. Survivorship analysis for cup loosening for any cause was 93% for group 1 (11 cups) and 99.2% for group 2 (1 cup in a patient with severe developmental dysplasia). The cumulative probability of not having a revision of one or both components for any cause was 96.9% (CI 95%:9.47–99.1%), All unrevised cases showed good clinical and radiographic results at the end of follow-up. No stem loosening, osteolysis nor stress shielding were found in any case by the end of follow-up. Changes in linear femoral head penetration were not seen in any hip. These data suggest that Cerafit alumina-on-alumina prostheses show excellent results after fiften years. Cup loosening was more frequent in the cups of first generation, but not currently. Continued follow-up will be required to determine if reduction in wear between the alumina-on alumina bearings results in less osteolysis and loosening.
Cup migration and bone graft resorption are some of the limitations after acetabular impaction bone grafting (IBG) technique in revision hip surgery when used for large segmental defects. We asked whether the use of a metallic mesh may decrease the appearance of this complication. We compared the appearance of loosening in patients with a bone defect 3A or 3B according to Paprosky. We assessed 204 hips operated with IBG and a cemented cup according to Slooff et al between 1997 and 2004. There were 100 hips with a preoperative bone defect of 3A and 104 with a 3B. We used 142 medial and/or rim metallic meshes for uncontained defects. The mean follow-up for unrevised cups was 10.4 years. We detemined postoperative radiological cup position and acetabular reconstruction of the hip center according to Ranawat in both groups. We assessed the appearance of radiological loosening and resorption of the graft.Background
Materials and Methods
Dislocation is one of the most important complications after primary total hip replacement (THR). The low incidence of this finding makes it difficult to analyse the possible risk factors. The surgical technique can also influence this rate through cup position or an adequate reconstruction of the hip. We assessed the demographic data and radiological reconstruction of the hip related to the appearance of dislocation after primary THR. 1414 uncemented THRs were recorded from our Local Joint Registry. The mean age of the patients was 60.1 years old (range, 14 to 95), and the mean weight was 73.3 kg (42 to 121). There were 733 men and 974 patients were classified with an activity level of 4 or 5 according to Devane. The most frequent diagnosis was primary osteoarthritis, 795 hips, followed by avascular necrosis 207 hips. An alumina-on-alumina THR was implanted in 703 hips and a metal-on-polyethylene THR in 711 hips. A femoral head size of 28 mm was used in 708 hips and 32 mm in 704. Radiological cup position was assessed using the acetabular abduction angle, the height of the center of the hip, and the horizontal distance of the cup. Cup anteversion was measured according to Widmer and the reconstruction of the center of rotation of the hip according to Ranawat. The radiographic reconstruction of the abductor mechanism was measured using two variables: the lever arm and the height of the greater trochanter.Introduction
Material and Methods
We report the results of 79 patients (81 hips)
who underwent impaction grafting at revision hip replacement using the
Exeter femoral stem. Their mean age was 64 years (31 to 83). According
to the Endoklinik classification, 20 hips had a type 2 bone defect,
40 had type 3, and 21 had type 4. The mean follow-up for unrevised
stems was 10.4 years (5 to 17). There were 12 re-operations due to intra- and post-operative
fractures, infection (one hip) and aseptic loosening (one hip).
All re-operations affected type 3 (6 hips) and 4 (6 hips) bone defects.
The survival rate for re-operation for any cause was 100% for type
2, 81.2% (95% confidence interval (CI) 67.1 to 95.3) for type 3,
and 70.8% (95% CI 51.1 to 90.5) for type 4 defects at 14 years.
The survival rate with further revision for aseptic loosening as
the end point was 98.6% (95% CI 95.8 to 100). The final clinical
score was higher for patients with type 2 bone defects than type
4 regarding pain, function and range of movement. Limp was most
frequent in the type 4 group (p <
0.001). The mean subsidence
of the stem was 2.3 mm ( The impacted bone grafting technique has good clinical results
in femoral revision. However, major bone defects affect clinical
outcome and also result in more operative complications.
Acetabular bone structure is not the same in all patients and can be defined by the radiolucent triangle superior to the acetabulum. We ask if the acetabular anatomy determines the initial cup fixation and screws use. We have assessed 205 hips in which a Cerafit cementless cup was implanted. According to Dorr et al., acetabulae were classified as type A, in which the radiolucent triangle had an isosceles shape (86 hips), type B, in which the triangle extended into the teardrop (90 hips), and type C which had a right-angle triangle (29 hips). The use of screws was decided at the time of surgery and according to cup stability, not acetabular anatomy. Avascular necrosis and inflammatory arthritis were the most frequent diagnoses in type A hips, osteoarthritis in type B, and dysplasia in type C. Women were more frequent in types A and C (p<
0.001). The use of screws was more frequent in women (p<
0.001) and in type A (34.9%) and type C hips (62.1%) than in type B hips (20.0%) (p<
0.001). The multivariate logistic regression model showed the acetabular type (p=0.11) and gender (p=0.003) as independent factors. Acetabular types A (OR=1.98, 95% CI: 0.922–4.208, p=0.075) and C (OR=5.09, 95% CI: 1.74–14.9, p=0.003) increase the risk for screw use. Men have a lower risk for screw use (OR=0.329, 95% CI: 0.16–0.68, p=0.003). Acetabular anatomy and gender determine the use of screws in cementless cups. Continued follow-up is necessary to determine if screws results in less loosening and osteolysis.
Revision surgery of the hip was performed on 114 hips using an extensively porous-coated femoral component. Of these, 95 hips (94 patients) had a mean follow-up of 10.2 years (5 to 17). No cortical struts were used and the cortical index and the femoral cortical width were measured at different levels. There were two revisions for aseptic loosening. Survivorship at 12 years for all causes of failure was 96.9% (95% confidence interval 93.5 to 100) in the best-case scenario. Fibrous or unstable fixation was associated with major bone defects. The cortical index (p = 0.045) and the lateral cortical thickness (p = 0.008) decreased at the proximal level over time while the medial cortex increased (p = 0.001) at the proximal and distal levels. An increase in the proximal medial cortex was found in patients with an extended transtrochanteric osteotomy (p = 0.026) and in those with components shorter than 25 cm (p = 0.008). The use of the extensively porous-coated femoral component can provide a solution for difficult cases in revision surgery. Radiological bony ingrowth is common. Although without clinical relevance at the end of follow-up, the thickness of the medial femoral cortex often increased while that of the lateral cortex decreased. In cases in which a shorter component was used and in those undertaken using an extended trochanteric osteotomy, there was a greater increase in thickness of the femoral cortex over time.
Cultures were polymicrobial in 22 cases and by Gram-positive in 55 (80.9%). Highly-resistant organisms: methicillin-resistant Staphylococcus (36 patients) and ESBL-producing Enterobacteriaceae (2 patients). “Problematic-treatment”: Enterococcus (6 patients), Pseudomonas (3 patients), non-fermenting Gram-negative (2), moulds (1). Oral antibiotic selection: according to bacterial sensitivity, biofilm and intracellular effectiveness. Protocolized surgery: two-stage exchange. Average follow-up: 4.7+/−2.7 years (1–11). Healing of infection is diagnosed if absence of clinical, serological and radiological signs of infection during the whole follow-up. Orthopaedic outcome is evaluated by HHS for hips and by KSCRS for knees.
Healing of infection: 59/68 patients (86.8%), 32/37 hips (86.5%) and 27/31 knees (87.1%). Infection not healed: 7/68 cases (10.3%) (4/37 hips, 3/31 knees) (5 by highly-resistant and 1 by “problematic-treatment” bacteria). There are no differences between hips and knees (p=0.55).
Statistically significant differences are not found when comparing subgroups according to Gram stain (p=0.43), multiple vs single bacteria (p=0.47 infective, p=0.71 orthopaedic), highly-resistant bacteria (p=0.2 infective, p=0.1/0.5 orthopaedic), or “problematic-treatment” (p=0.68).
A strong statistical correlation appears between infective and orthopedic results after late arthroplasty infections. With the number of cases presented significant differences in infective or in orthopaedic results are not found when comparing single vs. polymicrobial, gram-negative vs. gram-positive, high vs. low antimicrobial resistance and “problematic-treatment” infections.
Gestational age was 1st trimester (3 cases), 2nd trimester (5 cases), 3rd trimester (5 cases). 10 women were treated surgically, 8 before finishing gestation. Gestation ended as and induced abortion (3 cases, 1 due to fetal death and 2 due to teratogenic risk), and birth (10 cases, all alive, 50% eutocic). Only 3 babies needed type II or type III neonatal reanimation.
33 patients were treated from 1996 to 2002: 8 THA, 5 hip hemiarthroplasties, 20 TKA. Bacteriology: 24 Staphylococci of which 16 were methycillin-resistant, 7 multi-resistant Gram-negative, 2 Cory-nebacteriae; 7 polymicrobian. Antibiotic therapy: two simultaneous oral antibiotics, selected according to bacterial sensitivity and intracel-lular effectiveness (rifampin, ofloxacin, ciprofloxacin, levofloxacin, trimethoprim-sulfamethoxazole, fosfomicin, linezolid, doxiciclin), were used on an outpatient basis (between 1st and 2nd surgery, and after 2nd surgery until serological normalization). Patients received intravenous antibiotics and were in-hospital only for one week after surgery. Surgery: two-stage exchange with 2nd stage delayed until clinical and serological normalization. Healing of infection: absence of clinical, serological and radiological evidence of infection along all follow-up. Prospective follow-up: 24-96 months.
Treatment failure: 1 patient (TKA) (3%). THA: 8/8 infections healed: 1 Girdlestone patient (1st stage of exchange) rejected reimplantation; 7 two-stage exchange (good/excellent objective and subjective result). Hip hemiarthroplasty: 5/5 infections healed: 3 Girdlestone (1st stage of exchange surgery, 2nd stage rejected because of hemiplegia or Alzheimer); 2 two-stage exchange (good/excellent objective and subjective result). TKA: 19/20 infections healed: 3 resection-arthroplasty (1st stage of exchange surgery, 2nd stage rejected because of Buerger, cirrhosis or Alzheimer); 17 two-stage exchange (15 good/excellent objective and subjective results, 1 patient needed a debridement 2 months after 2nd surgery because of prolonged aseptic drainage and healed uneventfully, 1 failure described).
Between 1972 and 1990, we performed 168 primary low-friction arthroplasties in 125 patients with acetabular protrusion. Twelve hips were lost to follow-up within eight years and eight which became infected were excluded from the final study. Of the 148 hips remaining, 62 with a mild protrusion were classified as group 1, 54 with moderate or severe protrusion as group 2 and, after 1985, 32 with moderate and severe protrusion which required bone grafts as group 3. The mean follow-up was 18.3 years (3 to 24) for group 1, 17.4 years (8 to 22) for group 2 and ten years (8 to 13) for group 3. There were 31 revisions of the cup, 12 in group 1 and 19 in group 2. According to the Kaplan-Meier analysis the overall rates at 20 years were 21 ± 10.79% in group 1 and 37 ± 11.90% in group 2. There have been 43 radiological loosenings: 22 in group 1, 21 in group 2 and none so far in group 3, at ten years. The overall loosening rates at 20 years were 42 ± 14.76% in group 1 and 49 ± 19.50% in group 2. The grafts were well incorporated in all group-3 hips, and the bone structure appeared normal after one year. The distance between the centre of the head of the femoral prosthesis and the approximate true centre of the femoral head was less in group 3 than in groups 1 and 2 (p <
0.01). According to the Cox proportional-hazards regression this was the single most important factor in loosening of the cup (odds ratio 1.11; 95% CI 1.05 to 1.18/mm). Better results were obtained in moderate and severe protrusions reconstructed with bone grafting than in hips with mild protrusion which were not grafted.