Results of total hip replacement using the robodoc surgical assistant system: clinical outcome and evaluation of complications for 97 procedures

THE INTERNATIONAL JOURNAL OF MEDICAL ROBOTICS AND COMPUTER ASSISTED SURGERY ORIGINAL ARTICLE
Int J Med Robotics Comput Assist Surg 2007; 3: 301–306.
Published online 14 November 2007 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/rcs.161
Results of total hip replacement using the Robodoc
surgical assistant system: clinical outcome and
evaluation of complications for 97 procedures
Abstract
Background
A computerized robotic surgical system was developed from 1986 by the Thomas J. Watson Research Center. In 1992 the system unit Orthodoc and the milling robot Robodoc were first used on humans. Wepresent the results achieved with Robodoc-assisted total hip arthroplasty in Between 1997 and 2002, 143 total hip replacements (128 patients) were performed using the Robodoc system. This is a consecutiveseries. Complete follow-up was possible in 97 hips at a mean follow-up period Reconstructive Surgery, BG TraumaHospital, Hamburg, Germany Technical complications directly related to the robotic device occurred in nine cases (9.3%). The pre-operative Merle d’Aubigne score was determined at 8.1 points compared to a post-operative mean score of 16.2. There was no sign of femoral stem loosening on radiographs.
Conclusions
Robotic-assisted total hip arthroplasty with the Orthodoc/ Robodoc system achieves equal results as compared to a manual technique.
However, there was a high number of technical complications directly or indirectly related to the robot. Copyright  2007 John Wiley & Sons, Ltd.
Strasse 10, 21033 Hamburg,Germany. E-mail: Keywords
Introduction
To improve the results for permanent fixation in cementless total hip replace-ment (THR), a computerized robotic surgical system was developed from 1986by the Thomas J. Watson Research Center (IBM, USA) with Davis Universityin California (1). The company ISS (Integrated Surgical Systems, Sacramento,CA, USA) further developed a planning and robotic system, which consistedof the planning unit Orthodoc and the milling robot Robodoc (1). Afterclinical trials on dogs, in 1992 the system was first used on humans (1).
The first clinical use in Germany took place in the summer of 1994 at theBerufsgenossenschaftliche Unfallklinik (BGU) in Frankfurt (2). In the subse-quent 10 years the Robodoc system has sometimes received harsh criticismregarding the technique, results and complications caused by the system(3–10). Because of this, the French subsidiary, ISS SA (formerly IMMI) wasbrought into the French bankruptcy court and no further support was pro- vided to its users. ISS filed with the U.S. Securities and Exchange Commission Copyright  2007 John Wiley & Sons, Ltd.
A. P. Schulz et al.
(SEC), stating that it had ‘ceased operations and Patients had to be able to give informed consent to terminated all employees’. As a result, there are no systems remaining in clinical use in Europe. About 2 years later, • Exclusion criteria: age <18 years, revision THA, not able ISS resumed operations and subsequently sold its assets to Novatrix Biomedical in June 2007 (11). There werereports about high accuracy and stability achieved by The mean age at the index procedure was 56 (median the technique (5,12–14); on the other hand, conflicting 59, range 19–75) years. 89 hip replacements were results have been reported regarding the amount of soft implanted in male patients (62%). The reason for THA in tissue problems and gait anomalies found (9,14–16).
76 (53%) patients was idiopathic arthritis of the hip, in 43 In 1997 we introduced the Robodoc robotic system (30%) the arthritic deformity was due to a post-traumatic for total hip arthroplasty (THA) in our institution. The condition (mainly after fractured neck of femur in 23 rationale for the acquisition of such a system was first to (54%) cases). Eleven (8%) patients had dysplastic hips optimize the implant size and orientation, and second to and in 13 (9%) osteonecrosis of the femoral head was the optimize the accuracy of the created cavity. We hoped reason for THA (see Table 1). Pre-operative data included thereby to gain a significant increase in primary stability a general examination, measurement of range of motion, and an improvement in long-term results. Optimized determination of leg length inequality and radiographs.
osteo-integration of the implant was a major goal and In three cases (2.1%) the robotic operation had to the resulting load transfer from implant to bone was be stopped due to a technical fault of the robot and thought to have a decisive effect on the durability of the a conventional THA had to be performed; these patients prostheses. As a trauma hospital, we treat a comparatively where excluded. One patient developed a septic loosening high number of post-traumatic deformities, resulting in (Staphylococcus epidermidis) and had to be revised. At the need for THA. To monitor and analyse the results of follow-up, 23 patients (25 hips) could not be traced.
this completely new technique, we enrolled all patients Overall, 14 patients (15 hips) refused to take part in treated with the Robodoc for a clinical trial. The aim any further follow-up regarding this study. Two patients was to identify complications directly during the use died of unrelated causes. All these cases had to be of the robotic system as well as in the post-operative excluded from the study, leaving 85 patients (97 hips).
period. Retrospectively we then analysed these data The eventual follow-up rate is therefore 68% of implanted and performed follow-up examinations, including scoring There was no statistical difference (p > 0.05) in the As the Robodoc system is still in clinical use in India, group of all patients and the group that completed the South Korea and Japan and a FDA trial (17) is coming to study protocol. In the group of 97 patients with complete an end, we thought it of value to report the results gained follow-up, statistical differences were detectable in the with the assistance of the Orthodoc/Robodoc system.
group with post-traumatic arthritis compared to patientswith non-traumatic arthritis for the items age (50 vs.
Patients
Table 1. Aetiology of arthritis in all procedures and patients who
completed the study protocol
Between February 1997 and October 2002, 143 total hip replacements (128 patients) were performed using the Robodoc system. This was a consecutive series. No hip replacements using this system were performed after this time at our institution (see Figure 1).
• Inclusion criteria: considered for inclusion were all patients with an indication for cement-free THA.
Figure 1. Robotic-assisted THA performed per year over the study period (n = 143 hips)
Copyright  2007 John Wiley & Sons, Ltd.
Int J Med Robotics Comput Assist Surg 2007; 3: 301–306.
Evaluation of total hip replacement using the Robodoc system
61 years) and number of previous procedures (p < 0.05).
(pelvic view, 43 × 35 cm; axial Lauenstein projection, Although the rate of technical complications was 16% 20 × 40 cm) with comparison to direct post-operative in post-traumatic vs. 9% in non-traumatic arthrosis, the radiographs and examination of signs of osteolysis and difference was not statistically significant.
loosening under determination of the Gruen (20) andDeLee-Charnley (21) zones. The follow-up examinationand data exploration were performed by an independent examiner (a general practitioner, A.v.H) who was notinvolved in any stage of the treatment.
A modified transgluteal approach, as described by Bauer(18) was used in all cases. As a first step, two reference markers in the form of titanium pins were implanted atthe medial femoral condyle and the trochanter major, Data sampling of this study was prospective from the time respectively (the pinless version of the Robodoc system of admission. Before follow-up examination, the data was not used in this study). This was performed under were extracted to a database. Data from the follow-up local anaesthesia 1 day before the intended THA. After were added. The descriptive statistics were performed this, a CT scan, with approximately 80 slices of the using SPSS version 11.0 (SPSS Inc., Chicago, IL, USA).
proximal femur and 10 slices of the distal femur, was taken For statistical calculations, non-parametric methods were and transferred to the Orthodoc planning station. With applied (median, quartile). Means were compared with a colour-coded three-dimensional (3D) graphical display, non-parametric methods; for the median, the Wilcoxon the planned prostheses was chosen from an electronic catalogue and positioned by the surgeon. Positioning waspossible in all directions, including rotation; the accuracywas stated by ISS as 0.1 mm axially and 1◦ rotationally (19). The data were then transferred to the Robodoc.
After manual implantation of the acetabular compo- nent, the femur was fixated and the reference markers were digitized with the robotic system for matching with the CT data. The path to the femoral canal had to becompletely free of soft tissues to allow the required work- There were no anaesthesiological complications in this ing space for the robotic arm. Detection of significant series. The rate of technical complications was 9.3%, movement of the femur relative to the robot during the the rate of surgical complications 8.3% (for details, see procedure led to an immediate stop; the reference points Table 2). There was no statistical difference detectable had to be sought again and the milling had to be re- regarding patients with a post-traumatic or non-traumatic started. During the milling procedure, the milling process condition (p > 0.05).
could be visualized on a graphical display; milled boneparticles were removed using high-pressure irrigation.
After that, the femoral component was manually inserted and the procedure finished with placement of two wounddrainages and wound closure in anatomical layers. For During the pin implantation, one Kirschner wire broke, the acetabular component, the cement-free Osteoloc sys- one case of temporary lateral femoral cutaneal nerve tem (Howmedica, Rutherford, New Jersey, USA) was damage occurred and one post-operative knee effusion used in all cases. The femoral component consisted of a was seen (3.1% complication rate by the pin insertion cement-free implanted ABG II prosthesis (Howmedica) in process). In one case, the acetabular component was implanted at a suboptimal angle and had to be revised The surgeon filled out a sheet detailing the intra- during the procedure, due to a dislocation tendency. In operative problems and complications, divided by sur- one case, the acetabular reaming was performed too deep gical, technical and anaesthesiological issues. Single-shot and bone grafting was necessary as a consequence. Once, antibiotic prophylaxis (Cefuroxime i.v.) was used in all a femoral shaft fissure occurred during reposition and a cases; no antithrombotic pneumatic device was used intra- wire cerclage was necessary. Twice, intra-operative blood operatively. All patients received anti-thombosis stockingsand subcutaneous low-molecular weight heparin in the Table 2. Evaluation of intra-operative complications
post-operative period. Ossification prophylaxis consistedof twice-daily 50 mg indomethacin for 10 days. In 31 (32%) cases this had to be terminated, due to side-effects.
Follow-up examination took place on average 3.8 (range 2.9–5.3) years after the index procedure. The protocol for follow-up included a clinical examination, the Merle d’Aubigne Score–Postel hip score, radiographs Copyright  2007 John Wiley & Sons, Ltd.
Int J Med Robotics Comput Assist Surg 2007; 3: 301–306.
A. P. Schulz et al.
loss required the supplementation of two and three units, The pre-operative Merle d’Aubigne Postel score (22)was determined with a mean of 8.1 points compared to a post-operative mean score of 16.2 (see Table 3);the difference was statistically significant (p < 0.001).
There was a statistically significant difference between Technical complications directly related to the robotic the pre-operative score result of patients with a traumatic device occurred in nine cases (9.3%). Five times the condition and those with a non-traumatic condition milling process was halted by the bone motion monitor (p < 0.05), but the difference of the post-operative score and re-registration was necessary. Two femoral shaft result showed no statistical difference (p > 0.05).
fissures that required wire cerclage occurred duringmilling. Once the rim of the acetabulum was damagedby the milling device and once a defect at the greater Asked about their satisfaction with the procedure, 71patients (82 hips, 85%) stated that they were satisfied; 15 patients (15 hips, 15%) were dissatisfied. In the groupwith post-traumatic joint deformity, only 71% of patients Early post-operative complications (≤30 days post- were satisfied with the result, compared to 92% in the operatively) were noted in nine patients (9.3%). Three subgroup with a non-traumatic condition (p < 0.05).
times a significant haematoma developed; in two of thesecases the wound was revised. There was one dislocation12 days post-operatively (1%); after closed manipulation there was no recurrence. Twice a superficial infection wasnoted that healed under conservative measures; there was In 41 hip joints (42%), heterotopic ossifications were no deep infection in this series. There were three cases found (Table 4). One of these patients had previously had of deep venous thrombosis of the ipsilateral leg (twice an operation for the removal of symptomatic ossification below knee, once popliteal); one of those developed a (Brooker grade 3). Although Brooker grade 2 ossifications were found in four patients with a post-traumatic Two late complications included a symptomatic condition and in one case with a non-traumatic condition, Brooker type 3 heterotopic ossification that was removed there was no statistical difference in the subgroups with at 25 months post-operatively, and a painful scar that had traumatic or non-traumatic arthrosis (p > 0.05).
Evaluation of the femoral component showed an osteolysis in Gruen zones 1 and 4. There was no signof prosthesis loosening in any femoral component; allfemoral components were found to be adequately sized.
Results at follow-up
No component migration or subsidence was observed.
Table 3. Statistical evaluation of post-operative Merle d’ Aubigne
score (n = 97 hips)
A positive Trendelenburg sign was found in 17 patients (18%). The sign was positive in 12 patients (19.3%) with non-traumatic arthrosis and five patients (14%) with a post-traumatic condition. The difference was statistically not significant (p > 0.05). Leg length discrepancy of up to 10 mm was found in 19 patients (22%). A shortening of 10–30 mm was noted in 4/85 patients (4.7%); inthree of these the shortening was pre-existing, due to apost-traumatic condition, and in one case a pre-operative Table 4. Number and percentage of heterotopic ossifications
shortening of 1.5 cm had increased to a 3 cm deficiency.
according to the Brooker (23) classification at follow-up (n = 97)
Although this case occurred in the subgroup with a non-traumatic arthrosis, the difference was not significant (p > 0.05). The maximal flexion at follow-up was mean 107◦ (range 95–140◦); the mean extension was 0.8◦ extension deficit (10◦ deficit to 5◦ hyperextension). The abduction was determined at mean 27◦ (range 0–50◦).
There was no statistical difference between subgroups.
Copyright  2007 John Wiley & Sons, Ltd.
Int J Med Robotics Comput Assist Surg 2007; 3: 301–306.
Evaluation of total hip replacement using the Robodoc system
Peri-acetabular radiolucencies around the acetabular studies reporting the results of manual THA via the same component were noted in DeLee–Charnley zones at the approach, the reported rate is 13–19% in medium- to following rates: none in zone I, two in zone II, and one in long-term follow-up (32–36). We therefore can see no zone III. In addition, two implants demonstrated evidence evidence that robotic THA, at least with an anatomically- of migration. There was no sign of excessive polyethylene shaped stem, increases gait abnormalities compared to conventional implantation. This is supported by a studyanalysing gait patterns in patients after Robodoc THA andcomparing these to a control group with a conventional Discussion
The discrepancy in leg length that often exists in post- traumatic arthrosis can often not be equalized during hip The Robodoc system for robot-assisted total hip replace- arthroplasty. The same restrictions occur in robot-assisted ment was initially designed to increase the accuracy of procedures and left three patients with a leg shortening implantation by decreasing human errors (3). At the time of more than 1 cm; in one case the reason for this was of its development, it represented the most sophisticated faulty planning on the Orthodoc, a complication so far not piece of surgical equipment that the world of orthopaedics described in the literature. The dislocation rate of 1% is a had experienced to that date. By its use, osteointegration good result and in contrast to the 18% reported by Honl was thought to be optimized and a longer-lasting prosthe- et al. (14). In this study, the conceptually quite different sis seat was hoped to result. Evident advantages of this S-ROM prosthesis, which has a straight stem, was used, procedure were the possibility of pre-operative planning requiring an entirely different milling path and thereby and templating. Evident disadvantages were the increased possibly causing more soft tissue impairment (14).
theatre time, the direct cost increase and the necessity for The result as judged by the Merle d’Aubigne Postel a pre-operative CT scan. A further disadvantage of the (22) score was good, with an improvement from 8.1 system described here is the necessity for a further proce- points pre-operatively to 16.2 points at follow-up. This dure to implant two pins before the scan. The early results 8.1 point improvement is in the range of what has of robot-assisted THA showed no significant short-term been described before and after robotic THA (12,14,16).
benefit regarding mobilization, rehabilitation or clinical More importantly, there is no evidence that, at least result (9,14). Reports of a possible relationship between in short and medium outcomes, the results with the robotic THA and post-operative gait abnormalities led to Robodoc system are superior to those with manual great media coverage, especially in Germany (10), and a implantation techniques, with reported Merle d’Aubigne cessation of robot-assisted procedures in Westen Europe score improvements of 6.6–8.8 points (25,30,31).
The reported rates of loosening of the ABG II stem The rate of general orthopaedic complications is in an after manual implantation have so far been low (0–2%) area that is comparable to the described rates for manual on medium-term follow-up (25,30,31,37). It is thought implantation techniques (25–27). On the other hand, that this is due to improved early osseointegration the rate of technical complications caused by the robotic of this prosthesis model with its hydroxyapatite (HA) system was worrying. In 9.3% of procedures, technical coating (38). Equal results can be reached after robotic complications occurred. In 3.1%, these were caused by implantation, but there is no sign that results are the pins required for registration; these were not required better with a robotic technique. The evaluation of for the last available version of the Robodoc system.
osteolysis in the different Gruen zones (20) showed a low Although the application of these markers is often not amount of osteolysis as compared to manual implantation mentioned as a cause of complications (2,3,28), Nogler (25,30,31,37,39). It has been shown, however, that the reported about 10/18 patients having persistent severe amount of periprosthetic osteolysis is largely dependent pain at the site of pin implantation (7). Regarding the on the type and concept of prosthesis used (39,40); with rate of transformations from robotic to manual technique our data we are not in a position to state that robotic due to robot soft- and hardware failure, a rate of 2.1% milling decreases the amount of osteolysis. A rate of might seem high. Honl et al., however, had to switch trochanteric ossification of 27.8% is in the range of what procedures in over 18% of cases; in four of their cases, is also to be expected after manual implantation (41,42).
the reaming process did not start at all (14). An unusual We could see no evidence that the milling process, with its complication was the occurence of fissures during milling large amount of debris, increases the rate of ossifications.
in two cases (2%). So far this was thought not to occur in In conclusion, we were able to show that the the robotic milling process, and has not yet been reported results of robotic-assisted total hip arthroplasty with (2,14,28,29). This rate is about what can be expected in the Orthodoc/Robodoc system achieves equal results regarding functional outcome, radiographic outcome The prosthesis model used in this study was shown in and general complications as compared to a manual a computer simulation study to be favourable regarding technique. At the follow-up period reported, there is the extent of muscle detachment produced by robotic no evidence that at medium-term follow-up a robotic milling (14). We found a Trendelenburg sign in 18% of technique is superior in any of the areas examined. On patients after robotic THA at follow-up examination; in the other hand, there was a high frequency of technical Copyright  2007 John Wiley & Sons, Ltd.
Int J Med Robotics Comput Assist Surg 2007; 3: 301–306.
A. P. Schulz et al.
complications directly or indirectly related to the robot.
21. DeLee JG, Charnley J. Radiological demarcation of cemented The Robodoc had a very simple and plump milling device; total hip replacement. ClinOrthop 1976; 121: 20–32.
22. D’Aubigne RM, Postel M. Function al results of hip arthroplasty it is likely that this could have been improved if the with acrylic prosthesis. J Bone Joint Surg Am 1954; 36(A3):
development of the system had not halted. Future systems should address the complications of former systems and 23. Brooker AE, Bowerman JW, Robinson RA, Riley LH. Ectopic ossification following total hip replacement. Incidence and a the current trends in minimally invasive procedures.
method of classification. J Bone Joint Surg Am 1973; 55A:
1629–1632.
24. Schulz AP, Meiners J, Mantwill F, et al. Roboter in References
minimalinvasive Endoprothetik – das RomEo Projekt. Focus MUL
2006; 23(4): 182–187.
25. Giannikas KA, Din R, Sadiq S, Dunningham TH. Medium-term 1. Paul HA, Bargar WL, Mittlestadt B, et al. Development of a results of the ABG II total hip arthroplasty in young patients. J surgical robot for cementless total hip arthroplasty. Clin Orthop Arthroplasty 2002; 17(2): 184–188.
Relat Res 1992; 285: 57–66.
26. Perka C, Paul C, Matziolis G. [Factors influencing peri-operative 2. Borner M, Bauer A, Lahmer A. [Computer-assisted robotics in morbidity and mortality in primary hip arthroplasty]. Orthopade hip endoprosthesis implantation]. Unfallchirurg 1997; 100(8):
2004; 33(6): 715–720.
27. Sharkey PF, Shastri S, Teloken MA, et al. Relationship between 3. Birke A, Reichel H, Hein W, et al. [Robodoc – a path into the surgical volume and early outcomes of total hip arthroplasty: future of hip endoprosthetics or an investment error?]. Z Orthop do results continue to get better? J Arthroplasty 2004; 19(6):
Grenzgeb 2000; 138(5): 395–401.
4. Mantwill F, Schulz A, Faber A, et al. Robotic systems in total hip 28. Borner M, Bauer A, Lahmer A. [Computer-guided robot-assisted arthroplasty – is the time ripe for a new approach? Int J Med hip endoprosthesis]. Orthopade 1997; 26(3): 251–257.
Robotics Comput Assist Surg 2005; 1(4): 8–19.
29. Bargar WL, Bauer A, Borner M. Primary and revision total hip replacement using the Robodoc system. Clin Orthop Relat Res evaluation of femoral canal preparation using the Robodoc 1998; 354: 82–91.
system. J Orthop Sci 2004; 9(5): 452–461.
30. Rogers A, Kulkarni R, Downes EM. The ABG II hydroxyapatite- 6. Nogler M, Krismer M, Haid C, et al. Excessive heat generation coated hip prosthesis: 100 consecutive operations with during cutting of cement in the Robodoc hip-revision procedure.
Acta Orthop Scand 2001; 72(6): 595–599.
7. Nogler M, Maurer H, Wimmer C, et al. Knee pain caused by a 31. Tonino AJ, Rahmy AI. The hydroxyapatite–ABG II hip system: fiducial marker in the medial femoral condyle: a clinical and 5- to 7-year results from an international multicentre study. The anatomic study of 20 cases. Acta Orthop Scand 2001; 72(5):
International ABG II Study Group. J Arthroplasty 2000; 15(3):
8. Nogler M, Wimmer C, Lass-Florl C, et al. Contamination risk of the surgical team through Robodoc’s high-speed cutter. Clin outcome of total hip arthroplasty using the direct lateral vs Orthop Relat Res 2001; 387: 225–231.
the posterior surgical approach. Orthopedics 1996; 19(10):
9. Schrader P. [Consequence of evidence-based medicine and individual case appraisal of the Robodoc method for the MDK, 33. Pai VS. Significance of the Trendelenburg test in total hip and the malpractice management of insurance funds and the arthroplasty. Influence of lateral approaches. J Arthroplasty principles of managing innovations]. Gesundheitswesen 2005; 1996; 11(2): 174–179.
67(6): 389–395.
10. Englehardt R. [German patients as experimental rabbits. Spiegel primary total hip replacement. Int Orthop 1997; 21(6):
Magazine reports on implantation of hip endoprostheses with Robodoc]. Z Orthop Grenzgeb 2003; 141(1): 21.
11. Lamb C. ‘Robodoc’ creator sells assets to Irvine company.
arthroplasty: direct lateral or posterior? J Rheumatol 2004; Sacramento Business Journal 13 July 2007.
31(9): 1790–1796.
12. Nishihara S, Sugano N, Nishii T, et al. Comparison between 36. Jolles BM, Bogoch ER. Posterior vs. lateral surgical approach hand rasping and robotic milling for stem implantation in for total hip arthroplasty in adults with osteoarthritis. Cochrane cementless total hip arthroplasty. J Arthroplasty 2006; 21(7):
Database Syst Rev 2004; 1: CD003828.
13. Nishihara S, Sugano N, Nishii T, et al. Comparison of the fit hydroxyapatite-coated primary total hip arthroplasty: a clinical and fill between the anatomic hip femoral component and the and radiological review. J Orthop Surg (Hong Kong) 2003; 11(2):
VerSys Taper femoral component using virtual implantation on the Orthodoc workstation. J Orthop Sci 2003; 8(3): 352–360.
38. Garcia Araujo C, Fernandez Gonzalez J, Tonino A. Rheumatoid 14. Honl M, Dierk O, Gauck C, et al. Comparison of robotic-assisted arthritis and hydroxyapatite-coated hip prostheses: five-year and manual implantation of a primary total hip replacement.
results. International ABG II Study Group. J Arthroplasty 1998; A prospective study. J Bone Joint Surg Am 2003; 85(A8):
13(6): 660–667.
39. Rahmy AI, Gosens T, Blake GM, et al. Periprosthetic bone 15. Bach CM, Winter P, Nogler M, et al. No functional impairment remodelling of two types of uncemented femoral implant after Robodoc total hip arthroplasty: gait analysis in 25 patients.
with proximal hydroxyapatite coating: a 3-year follow-up study Acta Orthop Scand 2002; 73(4): 386–391.
addressing the influence of prosthesis design and pre-operative 16. Schrader P. [Technique evaluation for orthopaedic use of bone density on periprosthetic bone loss. Osteoporosis Int 2004; Robodoc]. Z Orthop Grenzgeb 2005; 143(3): 329–336.
15(4): 281–289.
17. Spencer EH. The Robodoc clinical trial: a robotic assistant for 40. Theis JC, Beadel G. Changes in proximal femoral bone mineral total hip arthroplasty. Orthop Nurs 1996; 15(1): 9–14.
density around a hydroxyapatite-coated hip joint arthroplasty. J Orthop Surg (Hong Kong) 2003; 11(1): 48–52.
transgluteal approach to the hip joint. Arch Orthop Trauma 41. Glas PY, Bejui-Hugues J, Carret JP. [Total hip arthroplasty after Surg 1979; 95(1–2): 47–49.
treatment of acetabular fracture]. Rev Chir Orthop Reparatrice 19. Product description: Robodoc/Orthodoc. Integrated Surgical Appar Mot 2005; 91(2): 124–131.
Systems, Sacramento, CA 95834, USA, 1997.
42. Harwin SF. Trochanteric heterotopic ossification after total 20. Gruen TA, McNeice GM, Amstutz HC. ‘Modes of failure’ of hip arthroplasty performed using a direct lateral approach. J cemented stem-type femoral components: a radiographic Arthroplasty 2005; 20(4): 467–472.
analysis of loosening. Clin Orthop Relat Res 1979; 141: 17–27.
Copyright  2007 John Wiley & Sons, Ltd.
Int J Med Robotics Comput Assist Surg 2007; 3: 301–306.

Source: http://www.robodoc.com.sg/references/32_Results_of_total_hip_replacement_using_the_Robodoc_surgical_assistant_system_clinical_outcome_and_evaluation_of_complications_for_97_procedures_2007_PN.pdf