- Email: [email protected]
Prognostic Factors Influencing the Functional Outcome of Total Hip Arthroplasty for Hip Infection Sequelae
The Journal of Arthroplasty Vol. 20 No. 5 2005
Prognostic Factors Influencing the Functional Outcome of Total Hip Arthroplasty for Hip Infection Sequelae Youn-Soo Park, MD, Young-Wan Moon, MD, Seung-Jae Lim, MD, Irvin Oh, MD, and Ji-Soon Lim, MD
Abstract: To investigate factors influencing functional prognosis for patients who have undergone total hip arthroplasty for late sequelae of infective arthritis of the hip, we conducted a retrospective analysis of 75 hips. All patients had had pyogenic or tuberculous infection and later underwent total hip arthroplasty. The average follow-up period was 5.8 years (range, 3-9 years). Various clinical and radiographic measures were analyzed with respect to Harris hip scores. Younger age at the time of infection onset, preoperative leg length discrepancy greater than 1 in, fusion, severe femoral hypoplasia, and severe acetabular dysplasia were associated with poorer prognosis. The complication rate, including 1 recurrent infection, was relatively low. Previous infections seemed to predict poorer results mostly because of the consequences of infection for the development of the hip joint, rather than because of infection recurrence. Key words: total hip arthroplasty, infective arthritis, infection sequelae, prognosis. n 2005 Elsevier Inc. All rights reserved.
hypoplasia of the femur. Although various special surgical techniques have been developed for THA in severely dysplastic hips in this subset of patients, relatively higher rates of complications, including intraoperative femoral fractures, recurrent infection, mechanical loosening, and revisions, have been reported [1-8]. However, some researchers have reported encouraging intermediate to long-term clinical results for THA in patients with childhood infection [3,5]. Moreover, both low rates of recurrence and no recurrence of infection have been reported [3,5,9,10]. Most studies have considered hips with pyogenic sequelae and hips with tuberculous sequelae separately, concentrating on the mechanical aspects of postsurgical complications and clinical results. However, there have been no reports, to the best of our knowledge, of clinical and radiographic factors that might serve as a predictor of functional success. We investigated which common clinical or radiographic factors correlated with the functional outcome of THA for infection sequelae in the hip joint.
Infective arthritis of the hip joint infrequently occurs in children and adults, and when it does, it is often in developing countries. If not treated promptly, it may yield various anatomic distortions of bony and soft-tissue structures, which may subsequently cause pain from secondary arthritis. Primary total hip arthroplasty (THA) in this patient group poses special problems, especially in the presence of scarred soft tissues, leg length discrepancy, severe flexion deformities, ankylosis, deficient bone stock, dysplasia of the acetabulum, or
From the Department of Orthopedic Surgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea. Submitted May 31, 2004; accepted March 31, 2005. No benefits or funds were received in support of the study. Reprint requests: Youn-Soo Park, MD, Department of Orthopedic Surgery, Samsung Medical Center, Ilwon-dong 50, Kangnam-ku, Seoul 135-710, Republic of Korea. n 2005 Elsevier Inc. All rights reserved. 0883-5403/05/1906-0004$30.00/0 doi:10.1016/j.arth.2005.04.003
608
Factors Influencing Outcome of THA for Hip Infection Sequelae ! Park et al 609
Materials and Methods Patient Selection and Diagnosis of Previous Infection Between December 1994 and March 2001, we performed THA in 97 patients whose history and clinical and radiographic findings suggested late infection sequelae of the hip joint. Diagnostic criteria included chronic hip pain in patients with clear evidence of previous infection of the hip joint (eg, medical or surgical treatment) and radiographic or clinical features of previous infection (eg, the presence of an arthrotomy scar or sinus tract). For all 97 patients, preoperative evaluation of persistent infection consisted of obtaining a complete blood count and sedimentation rate. For patients presenting with suspicious ongoing infection, preoperative hip joint aspiration was also performed. The average follow-up period after surgery was 5.8 years (range, 3-9 years). Six patients were lost to follow-up monitoring for at least 3 years; their data were excluded from the study. Of the remaining 91 hips, a cemented stem was implanted in 16 and a cementless stem in 75. Hips with cemented stems were excluded from the study to eliminate the influence of difference in fixation methods. Thus, clinical and radiographic information for 75 hips (in 75 patients: 39 women and 36 men) formed the basis of our retrospective analysis. The mean age of patients at the time of infection onset was 12 years (range, 1 month to 40 years), and mean age of patients at the time of the index operation was 51 years (range, 24 -75 years). The average time interval between initial infection contracture and primary arthroplasty was 34 years (range, 11-61 years). We operated only on patients who had not had any sign or symptom of infection for 10 years before the index operation, as suggested by Kim [5]. We obtained intraoperative swab and tissue cultures to investigate growth of aerobic, anaerobic, and tubercle bacilli. We also conducted histologic evaluation to detect any chronic or active infection. After surgery, we administered cephalosporin intravenously for 3 days, then orally for 1 week. Analysis of patients’ history of infection and radiographic and physical findings indicated that hip pain was likely caused by tuberculous infection sequelae in 26 patients and by pyogenic infection sequelae in 49 patients (Table 1). Clinical Variables For each patient, range of hip motion was recorded before and after surgery. The arc of motion was assessed as the angular difference of
Table 1. Summary of Patient Demographics Infection type Number of patients Sex (male/female) Age at onset of infection (range) Age at time of operation (range) Quiescent period (range) Previous treatment Arthrotomy and medication Medication only None Draining sinus scar (%)
Tuberculous infection
Pyogenic infection
26 13/13 17.6 y (1-40 y)
49 23/26 10.1 y (0.1-37 y)
52.8 y (35-75 y)
50.8 y (24-73 y)
31.4 y (11-61 y)
36.3 y (16-61 y)
25
41
0 1 7 (26.9)
1 7 24 (49.0)
further flexion and flexion contracture. Leg length discrepancy, which was measured as the difference in the distance from anterosuperior iliac spine to medial malleolus in affected and unaffected sides, was also recorded before and after surgery. There were 12 fused hips. A cementless stem was implanted in all 75 patients: 58 porouscoated S-ROM modular stems (DePuy, Warsaw, Ind), 6 Wagner conical stems (Protek, Berne, Switzerland), 10 C2 stems (Lima, Villanova, Italy), and 1 hydroxyapatite-coated Omnifit stem (Osteonics, Allendale, NJ). Subtrochanteric femoral shortening was performed in 3 hips. In all hips, a porous-coated, metal-backed modular cementless cup with varying size (range, 38-56 mm) was used. In 6 hips with a relatively high degree of acetabular dysplasia and deficient bone stock, we did excessive medial wall reaming and used a smaller cup with a 22-mm-diameter metal head and a polyethylene liner. The other 69 hips were implanted with a cup with a 28-mm metal head and a polyethylene liner. Two acetabular components were fixed with structural autogenous bone graft and screws without cement. We attempted to restore the normal hip joint center and to equalize leg length difference in all patients. Radiographic Findings We completed preoperative radiographic assessment of anatomic deformation for each hip. Using the system of Crowe et al [11], we classified patients into 4 groups based on degree of hip dysplasia: group 1, less than 50% subluxation; group 2, 50% to 75% subluxation; group 3, 75% to 100% subluxation; and group 4, more than 100% subluxation. To evaluate the degree of proximal
610 The Journal of Arthroplasty Vol. 20 No. 5 August 2005 recent follow-up HHS. The Mann-Whitney U test, analysis of variance multiple comparison test (the least significant difference test), and Spearman correlation analysis were performed using standard software (SPSS for Windows, Version 11.0, SPSS, Chicago, Ill), and P values of less than .05 were considered statistically significant. Surgical Technique
Fig. 1. Assessment of hypoplasia of the affected femur by measuring the cortical width of the isthmus in both sides. The relative difference of the isthmus width of the affected side with respect to the normal side was expressed as a percentage by calculating [(A B)/A] 100.
femoral dysplasia in plain anteroposterior radiograph, we assessed the relative hypoplasia of the affected femoral isthmus by measuring the difference in cortical widths of affected and unaffected femoral canals at the isthmus, expressed with respect to a normal isthmus cortical width as a percentage (Fig. 1). Harris Hip Score The Harris hip score (HHS) was determined before surgery and at each designated follow-up visit [12]. The average preoperative and last followup HHSs were 62.4 (range, 41-79) and 91.1 (range, 74-98), respectively.
Most patients underwent surgery in the lateral decubitus position, with a curvilinear lateral incision for anterolateral exposure. In fused or highly dislocated hips, we used the Smith-Peterson approach. We performed a trochanteric osteotomy in 6 severely dysplastic hips. The hip joint capsule was exposed by dissection of the anterior bundle of the gluteus medius and minimus muscles. The capsule was incised, and the femoral head was exposed and dislocated for osteotomy. The joint capsule was dissected and followed to reach the true acetabulum. In all patients, total excision of the capsule, followed by iliopsoas tenotomy, was carried out. If there was persistent flexor deformity, rectus and sartorius (flexor) recession was also performed. When there was rotational contracture, short external rotator release was performed, followed by gluteus maximus tenotomy. For 3 exceptionally high-riding femurs, which were difficult to relocate, we performed a femoral shortening procedure (Fig. 2).
Results Type of Infection and Age at Infection Onset
Statistical Analysis We determined the statistical significance of each clinical and radiographic factor using the most
There was no significant difference between the results for the pyogenic and tuberculosis groups, which showed average HHSs of 91.2 (range, 74 -98)
Fig. 2. A, Preoperative radiograph of the hip of a 44-year-old woman who had had infective hip arthritis at the age of 3 years. B, Radiograph obtained 3 years after surgery, which involved adductor tenotomy, femoral shortening osteotomy, and THA with a 22-mm head.
Factors Influencing Outcome of THA for Hip Infection Sequelae ! Park et al 611
and 91.3 (range, 83-98), respectively ( P = .647). Twenty-one patients who acquired an infection at 5 years or younger had an average final follow-up HHS of 88.7 (range, 75-98), whereas 54 patients who acquired an infection after 5 years of age had an average final follow-up HHS of 92.2 (range, 74 -98). The difference in scores between the 2 groups was statistically significant ( P = .008). Flexion Contracture and Fusion of the Hip Patients had an average preoperative flexion contracture of 19.88 (range, 08-608). The relation of preoperative flexion contracture to the final follow-up HHS was not statistically significant ( P = .456; Spearman correlation analysis). The preoperative and postoperative average arcs of motion (further flexion-flexion contracture) were 49.98 (range, 08-1258) and 93.58 (range, 608-1208), respectively. The average improvement in arc of motion after surgery was 46.98 (range, 358-1108), which was not significantly related to HHS ( P = .965). The average final follow-up HHS for 12 patients who had a fused hip before surgery was 87.8 (range, 80-96), whereas the average final follow-up HHS for 63 patients without a fused hip was 91.8 (range, 74 -98) —a statistically significant difference ( P = .005). These data suggest that the degree of preoperative flexion contracture did not significantly influence the clinical outcome, whereas fusion status was associated with a less favorable clinical outcome (Fig. 3).
Leg Length Difference The average initial leg length difference for the 75 patients was 3.11 cm (range, 0-8.5 cm). Thirtytwo patients whose initial leg length difference was less than 1 in (b2.54 cm) had an average final follow-up HHS of 93.8 (range, 85-98), whereas 43 patients whose initial leg length difference was 1 in or more had an average final follow-up HHS of 89.2 (range, 74 -96) — a significant difference ( P b .001). Acetabular Dysplasia There were 32 cases of Crowe group 1, 25 cases of Crowe group 2, 10 cases of Crowe group 3, and 8 cases of Crowe group 4 dysplasia of the hip. The average HHSs for each group at final follow-up examination were 93.3 (range, 85-98), 91.4 (range, 80-98), 90.1 (range, 85-95), and 83.1 (range, 74 -90), respectively. The analysis of variance multiple comparison analysis showed no statistically significant difference between groups 1, 2, and 3 ( P N .05) and a significant difference between group 4 and the other 3 groups ( P = .001). Femoral Hypoplasia The difference of cortical width at the isthmus of affected femurs vs that of unaffected femurs was assessed and expressed as a percentage of the width of the diameter of the unaffected isthmus. The average difference in the cortical widths of both isthmi with respect to the width of a normal isthmus was 12.9% (range, 0%- 49.2%). Fifty-nine patients had a cortical width difference of less than 20%, and 16 had a difference of more than 20%. There was a clinically significant difference between the 2 groups’ average final follow-up HHSs: 92.1 (range, 80-98) for the former group and 87.8 (range, 74-96) for the latter group ( P = .024). Reactivation of Infection
Fig. 3. A, Preoperative radiograph of the hip of a 38-year-old woman who had pyogenic hip arthritis at 4 years of age and developed fused hip. B, Radiograph obtained 6 years after THA with a modular stem shows satisfactory fixation of components. However, the patient reports discomfort due to residual leg length difference.
All intraoperative swab and tissue cultures were negative for tuberculous or bacterial infection, and only 3 specimens showed histologic signs of chronic inflammation. The remainder of the specimens showed no presence of any organism and no signs that might suggest acute or chronic inflammation. In 1 of the 3 patients, mildly inflamed hip joint fluid was found in preoperative aspirate, but the patient’s blood count and sedimentation rate were normal, and there was no sign of ongoing inflammation. The patient did not develop recurrent infection after the operation. Infection recurred
612 The Journal of Arthroplasty Vol. 20 No. 5 August 2005 after THA in 1 patient (1.3%) who had had tuberculous hip arthritis; the patient subsequently underwent a Girdlestone hip operation. The interval between active infection and arthroplasty in this patient was 11 years, and the patient’s preoperative sedimentation rate was normal. Radiographic Results One active young patient underwent cup revision because of polyethylene wear and loosening 6 years after the index operation. Three hips showed radiographic evidence of a loose acetabular component. Of those, 2 hips that had required structural autograft and screw fixations for severe acetabular dysplasia developed loosening and subsequently underwent revision. No radiographically loose femoral stem was detected at the final followup examination. Thus, the overall mechanical loosening rate was 5.3%. All of the failed acetabular components were in patients with Crowe group 4 hip dysplasia. Complications One female patient developed sciatic nerve palsy, from which she partially recovered, with mild sensory decrement, by 4 years after surgery. One periprosthetic fracture occurred at the anterior cortex of stem tip in a patient who underwent stem implantation after femoral shortening. The patient was successfully treated with open reduction and internal fixation 2 weeks after the index operation.
Discussion Our study concentrated on prognostic factors for THA performed because of infection sequelae, regardless of type. Because tuberculosis is endemic to Korea, we often found it difficult to clearly distinguish old tuberculous sequelae from pyogenic sequelae of the hip. Both types of infections are frequently encountered in Asian countries and tend to yield similar deformity. Although they may manifest different clinical features at the time of active infection, both types show similar softtissue scarring and destruction of bony structures. Clinical results in this specific group of patients were mainly influenced by age at the time of onset of the infection and extent of preoperative bony deformations (as represented clinically by large leg length difference or fusion of the hip, and radiographically by severe acetabular dysplasia with or without hypoplastic femur). Neither
preoperative flexion contracture nor improved arc of motion after surgery, however, influenced clinical results. Patients’ age at the time of infection onset was related to clinical results. Patients who acquired an infection at 5 years or younger had a lower HHS at final follow-up evaluation, most likely because the infection impeded skeletal growth and the development of anatomic structures, causing a greater degree of deformation. Although we attempted to correct leg length difference in all cases, some discrepancy inadvertently remained in some patients with a relatively larger degree of leg length inequality. When the initial leg length difference was more than 1 in, there was a relatively higher chance that some deformity would remain after surgery and produce a less favorable clinical outcome. Patients with severe acetabular dysplasia (Crowe group 4) had less-than-optimal clinical outcomes, with an average final follow-up HHS of 83.1 (range, 74-90). Four hips (5.3%) had aseptic loosening of the acetabular component, 3 of them subsequently undergoing revision. Of those 3 hips, 2 required structural autograft and screw fixation for severe acetabular dysplasia at the time of primary arthroplasty. Severe anatomic deformation of the acetabulum can be managed with a Charnley cemented cup or with bulk autogenous graft or allograft from the femoral head and insertion of the acetabular component with cement, but the reported results have not been encouraging for either fixation method [13,14]. Anderson and Harris [4], in their intermediateterm follow-up study, reported that stable fixation was achieved with the use of cementless acetabular components fixed with screws in hips with severe acetabular dysplasia, provided that at least 70% of the porous surface was covered by host bone. Because we have had experience with acetabular failure with structural autograft, we implanted 22-mm-head components and small (38-mm, 40-mm, and 42-mm) acetabular cups with medialization to gain better bony coverage in hips with severe acetabular dysplasia. Hypoplastic femur with a small proximal diameter secondary to undergrowth or destruction may limit the size and shape of the stem. Some authors have reported a high rate of revision due to mismatch of stem size with the canal. Kim found a high rate of loosening (17%) with the Mqller CDH femoral prosthesis (Protek) and attributed it to use of small femoral stems in normal femoral canals. Intraoperative fractures have been reported, and it has been suggested that poor initial
Factors Influencing Outcome of THA for Hip Infection Sequelae ! Park et al 613
stability is related to subsequent loosening [5]. Some authors advocate use of subtrochanteric shortening with derotational osteotomy or of a specially designed prosthesis, such as a CDH stem with swan-neck femoral components [1-7]. Such techniques have reportedly improved clinical outcomes in midterm to long-term studies. In our study, with an average follow-up period of 5.8 years, we found no radiologically loose stems. We expect that the loosening rate will remain low in the long term, considering the characteristics of porous-coated implants, which tend to produce persistent initial fixation. We believe that despite the femoral hypoplasia we encountered, the development of special stems, such as a modular design with various sizes, and current surgical techniques were associated with our improved clinical results. In our investigation, recurrence of infection after THA in this group of patients was uncommon; however, functional outcome was compromised by the extent of anatomic deformation caused by the initial infection. Age at the time of infection onset, preoperative leg length difference, presence of hip fusion, presence of severe dysplastic acetabulum as shown on radiographs, and the presence of proximal femoral hypoplasia (reflected in a small femoral isthmus cortical width) may be predictive of functional outcome.
References 1. DiFazio F, Shon WY, Salvati EA, et al. Long-term results of total hip arthroplasty with a cemented custom-designed swan-neck femoral component for congenital dislocation or severe dysplasia: a followup note. J Bone Joint Surg Am 2002;84:204. 2. Xener N, Tfzqn I˙R, AYik M. Femoral shortening and cementless arthroplasty in high congenital dislocation of the hip. J Arthroplasty 2002;17:41.
3. Kim YH. Total hip arthroplasty in adult patients who had childhood infection of the hip. J Bone Joint Surg Am 2003;85:198. 4. Anderson MJ, Harris WH. Total hip arthroplasty with insertion of the acetabular component without cement in hips with total congenital dislocation or marked congenital dysplasia. J Bone Joint Surg Am 1999;81:347. 5. Kim YH. Total arthroplasty of the hip after childhood sepsis. J Bone Joint Surg Br 1991;73:783. 6. Klisic P, Jankovic L. Combined procedure of open reduction and shortening of the femur in treatment of congenital dislocation of the hips in older children. Clin Orthop 1976;119:60. 7. Masonis JL, Patel JV, Miu A, et al. Subtrochanteric shortening and derotational osteotomy in primary total hip arthroplasty for patients with severe hip dysplasia. J Arthroplasty 2003;18 Suppl 1:68. 8. Cherney DL, Amstutz HC. Total hip replacement in the previously septic hip. J Bone Joint Surg Am 1983;65:1256. 9. Hardinge K, Cleary J, Charnley J. Low-friction arthroplasty for healed septic and tuberculous arthritis. J Bone Joint Surg Br 1979;61:144. 10. Jupiter JB, Karchmer AW, Lowell JD, et al. Total hip arthroplasty in the treatment of adult hips with current or quiescent sepsis. J Bone Joint Surg Am 1981;63:194. 11. Crowe JF, Mani VJ, Ranawat CS. Total hip replacement in congenital dislocation and dysplasia of the hip. J Bone Joint Surg 1979;61:15. 12. Harris WH. Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. J Bone Joint Surg Am 1969; 51:737. 13. Gerber SD, Harris WH. Femoral head autografting to augment acetabular deficiency in patients requiring total hip replacement: a minimum five-year and an average seven-year follow-up study. J Bone Joint Surg Am 1986;68:1241. 14. Numair J, Joshi AB, Murphy KC, et al. Total hip arthroplasty for congenital dysplasia or dislocation of the hip: survivorship analysis and long-term results. J Bone Joint Surg Am 1997;79:1352.
Prognostic Factors Influencing the Functional Outcome of Total Hip Arthroplasty for Hip Infection Sequelae Youn-Soo Park, MD, Young-Wan Moon, MD, Seung-Jae Lim, MD, Irvin Oh, MD, and Ji-Soon Lim, MD
Abstract: To investigate factors influencing functional prognosis for patients who have undergone total hip arthroplasty for late sequelae of infective arthritis of the hip, we conducted a retrospective analysis of 75 hips. All patients had had pyogenic or tuberculous infection and later underwent total hip arthroplasty. The average follow-up period was 5.8 years (range, 3-9 years). Various clinical and radiographic measures were analyzed with respect to Harris hip scores. Younger age at the time of infection onset, preoperative leg length discrepancy greater than 1 in, fusion, severe femoral hypoplasia, and severe acetabular dysplasia were associated with poorer prognosis. The complication rate, including 1 recurrent infection, was relatively low. Previous infections seemed to predict poorer results mostly because of the consequences of infection for the development of the hip joint, rather than because of infection recurrence. Key words: total hip arthroplasty, infective arthritis, infection sequelae, prognosis. n 2005 Elsevier Inc. All rights reserved.
hypoplasia of the femur. Although various special surgical techniques have been developed for THA in severely dysplastic hips in this subset of patients, relatively higher rates of complications, including intraoperative femoral fractures, recurrent infection, mechanical loosening, and revisions, have been reported [1-8]. However, some researchers have reported encouraging intermediate to long-term clinical results for THA in patients with childhood infection [3,5]. Moreover, both low rates of recurrence and no recurrence of infection have been reported [3,5,9,10]. Most studies have considered hips with pyogenic sequelae and hips with tuberculous sequelae separately, concentrating on the mechanical aspects of postsurgical complications and clinical results. However, there have been no reports, to the best of our knowledge, of clinical and radiographic factors that might serve as a predictor of functional success. We investigated which common clinical or radiographic factors correlated with the functional outcome of THA for infection sequelae in the hip joint.
Infective arthritis of the hip joint infrequently occurs in children and adults, and when it does, it is often in developing countries. If not treated promptly, it may yield various anatomic distortions of bony and soft-tissue structures, which may subsequently cause pain from secondary arthritis. Primary total hip arthroplasty (THA) in this patient group poses special problems, especially in the presence of scarred soft tissues, leg length discrepancy, severe flexion deformities, ankylosis, deficient bone stock, dysplasia of the acetabulum, or
From the Department of Orthopedic Surgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea. Submitted May 31, 2004; accepted March 31, 2005. No benefits or funds were received in support of the study. Reprint requests: Youn-Soo Park, MD, Department of Orthopedic Surgery, Samsung Medical Center, Ilwon-dong 50, Kangnam-ku, Seoul 135-710, Republic of Korea. n 2005 Elsevier Inc. All rights reserved. 0883-5403/05/1906-0004$30.00/0 doi:10.1016/j.arth.2005.04.003
608
Factors Influencing Outcome of THA for Hip Infection Sequelae ! Park et al 609
Materials and Methods Patient Selection and Diagnosis of Previous Infection Between December 1994 and March 2001, we performed THA in 97 patients whose history and clinical and radiographic findings suggested late infection sequelae of the hip joint. Diagnostic criteria included chronic hip pain in patients with clear evidence of previous infection of the hip joint (eg, medical or surgical treatment) and radiographic or clinical features of previous infection (eg, the presence of an arthrotomy scar or sinus tract). For all 97 patients, preoperative evaluation of persistent infection consisted of obtaining a complete blood count and sedimentation rate. For patients presenting with suspicious ongoing infection, preoperative hip joint aspiration was also performed. The average follow-up period after surgery was 5.8 years (range, 3-9 years). Six patients were lost to follow-up monitoring for at least 3 years; their data were excluded from the study. Of the remaining 91 hips, a cemented stem was implanted in 16 and a cementless stem in 75. Hips with cemented stems were excluded from the study to eliminate the influence of difference in fixation methods. Thus, clinical and radiographic information for 75 hips (in 75 patients: 39 women and 36 men) formed the basis of our retrospective analysis. The mean age of patients at the time of infection onset was 12 years (range, 1 month to 40 years), and mean age of patients at the time of the index operation was 51 years (range, 24 -75 years). The average time interval between initial infection contracture and primary arthroplasty was 34 years (range, 11-61 years). We operated only on patients who had not had any sign or symptom of infection for 10 years before the index operation, as suggested by Kim [5]. We obtained intraoperative swab and tissue cultures to investigate growth of aerobic, anaerobic, and tubercle bacilli. We also conducted histologic evaluation to detect any chronic or active infection. After surgery, we administered cephalosporin intravenously for 3 days, then orally for 1 week. Analysis of patients’ history of infection and radiographic and physical findings indicated that hip pain was likely caused by tuberculous infection sequelae in 26 patients and by pyogenic infection sequelae in 49 patients (Table 1). Clinical Variables For each patient, range of hip motion was recorded before and after surgery. The arc of motion was assessed as the angular difference of
Table 1. Summary of Patient Demographics Infection type Number of patients Sex (male/female) Age at onset of infection (range) Age at time of operation (range) Quiescent period (range) Previous treatment Arthrotomy and medication Medication only None Draining sinus scar (%)
Tuberculous infection
Pyogenic infection
26 13/13 17.6 y (1-40 y)
49 23/26 10.1 y (0.1-37 y)
52.8 y (35-75 y)
50.8 y (24-73 y)
31.4 y (11-61 y)
36.3 y (16-61 y)
25
41
0 1 7 (26.9)
1 7 24 (49.0)
further flexion and flexion contracture. Leg length discrepancy, which was measured as the difference in the distance from anterosuperior iliac spine to medial malleolus in affected and unaffected sides, was also recorded before and after surgery. There were 12 fused hips. A cementless stem was implanted in all 75 patients: 58 porouscoated S-ROM modular stems (DePuy, Warsaw, Ind), 6 Wagner conical stems (Protek, Berne, Switzerland), 10 C2 stems (Lima, Villanova, Italy), and 1 hydroxyapatite-coated Omnifit stem (Osteonics, Allendale, NJ). Subtrochanteric femoral shortening was performed in 3 hips. In all hips, a porous-coated, metal-backed modular cementless cup with varying size (range, 38-56 mm) was used. In 6 hips with a relatively high degree of acetabular dysplasia and deficient bone stock, we did excessive medial wall reaming and used a smaller cup with a 22-mm-diameter metal head and a polyethylene liner. The other 69 hips were implanted with a cup with a 28-mm metal head and a polyethylene liner. Two acetabular components were fixed with structural autogenous bone graft and screws without cement. We attempted to restore the normal hip joint center and to equalize leg length difference in all patients. Radiographic Findings We completed preoperative radiographic assessment of anatomic deformation for each hip. Using the system of Crowe et al [11], we classified patients into 4 groups based on degree of hip dysplasia: group 1, less than 50% subluxation; group 2, 50% to 75% subluxation; group 3, 75% to 100% subluxation; and group 4, more than 100% subluxation. To evaluate the degree of proximal
610 The Journal of Arthroplasty Vol. 20 No. 5 August 2005 recent follow-up HHS. The Mann-Whitney U test, analysis of variance multiple comparison test (the least significant difference test), and Spearman correlation analysis were performed using standard software (SPSS for Windows, Version 11.0, SPSS, Chicago, Ill), and P values of less than .05 were considered statistically significant. Surgical Technique
Fig. 1. Assessment of hypoplasia of the affected femur by measuring the cortical width of the isthmus in both sides. The relative difference of the isthmus width of the affected side with respect to the normal side was expressed as a percentage by calculating [(A B)/A] 100.
femoral dysplasia in plain anteroposterior radiograph, we assessed the relative hypoplasia of the affected femoral isthmus by measuring the difference in cortical widths of affected and unaffected femoral canals at the isthmus, expressed with respect to a normal isthmus cortical width as a percentage (Fig. 1). Harris Hip Score The Harris hip score (HHS) was determined before surgery and at each designated follow-up visit [12]. The average preoperative and last followup HHSs were 62.4 (range, 41-79) and 91.1 (range, 74-98), respectively.
Most patients underwent surgery in the lateral decubitus position, with a curvilinear lateral incision for anterolateral exposure. In fused or highly dislocated hips, we used the Smith-Peterson approach. We performed a trochanteric osteotomy in 6 severely dysplastic hips. The hip joint capsule was exposed by dissection of the anterior bundle of the gluteus medius and minimus muscles. The capsule was incised, and the femoral head was exposed and dislocated for osteotomy. The joint capsule was dissected and followed to reach the true acetabulum. In all patients, total excision of the capsule, followed by iliopsoas tenotomy, was carried out. If there was persistent flexor deformity, rectus and sartorius (flexor) recession was also performed. When there was rotational contracture, short external rotator release was performed, followed by gluteus maximus tenotomy. For 3 exceptionally high-riding femurs, which were difficult to relocate, we performed a femoral shortening procedure (Fig. 2).
Results Type of Infection and Age at Infection Onset
Statistical Analysis We determined the statistical significance of each clinical and radiographic factor using the most
There was no significant difference between the results for the pyogenic and tuberculosis groups, which showed average HHSs of 91.2 (range, 74 -98)
Fig. 2. A, Preoperative radiograph of the hip of a 44-year-old woman who had had infective hip arthritis at the age of 3 years. B, Radiograph obtained 3 years after surgery, which involved adductor tenotomy, femoral shortening osteotomy, and THA with a 22-mm head.
Factors Influencing Outcome of THA for Hip Infection Sequelae ! Park et al 611
and 91.3 (range, 83-98), respectively ( P = .647). Twenty-one patients who acquired an infection at 5 years or younger had an average final follow-up HHS of 88.7 (range, 75-98), whereas 54 patients who acquired an infection after 5 years of age had an average final follow-up HHS of 92.2 (range, 74 -98). The difference in scores between the 2 groups was statistically significant ( P = .008). Flexion Contracture and Fusion of the Hip Patients had an average preoperative flexion contracture of 19.88 (range, 08-608). The relation of preoperative flexion contracture to the final follow-up HHS was not statistically significant ( P = .456; Spearman correlation analysis). The preoperative and postoperative average arcs of motion (further flexion-flexion contracture) were 49.98 (range, 08-1258) and 93.58 (range, 608-1208), respectively. The average improvement in arc of motion after surgery was 46.98 (range, 358-1108), which was not significantly related to HHS ( P = .965). The average final follow-up HHS for 12 patients who had a fused hip before surgery was 87.8 (range, 80-96), whereas the average final follow-up HHS for 63 patients without a fused hip was 91.8 (range, 74 -98) —a statistically significant difference ( P = .005). These data suggest that the degree of preoperative flexion contracture did not significantly influence the clinical outcome, whereas fusion status was associated with a less favorable clinical outcome (Fig. 3).
Leg Length Difference The average initial leg length difference for the 75 patients was 3.11 cm (range, 0-8.5 cm). Thirtytwo patients whose initial leg length difference was less than 1 in (b2.54 cm) had an average final follow-up HHS of 93.8 (range, 85-98), whereas 43 patients whose initial leg length difference was 1 in or more had an average final follow-up HHS of 89.2 (range, 74 -96) — a significant difference ( P b .001). Acetabular Dysplasia There were 32 cases of Crowe group 1, 25 cases of Crowe group 2, 10 cases of Crowe group 3, and 8 cases of Crowe group 4 dysplasia of the hip. The average HHSs for each group at final follow-up examination were 93.3 (range, 85-98), 91.4 (range, 80-98), 90.1 (range, 85-95), and 83.1 (range, 74 -90), respectively. The analysis of variance multiple comparison analysis showed no statistically significant difference between groups 1, 2, and 3 ( P N .05) and a significant difference between group 4 and the other 3 groups ( P = .001). Femoral Hypoplasia The difference of cortical width at the isthmus of affected femurs vs that of unaffected femurs was assessed and expressed as a percentage of the width of the diameter of the unaffected isthmus. The average difference in the cortical widths of both isthmi with respect to the width of a normal isthmus was 12.9% (range, 0%- 49.2%). Fifty-nine patients had a cortical width difference of less than 20%, and 16 had a difference of more than 20%. There was a clinically significant difference between the 2 groups’ average final follow-up HHSs: 92.1 (range, 80-98) for the former group and 87.8 (range, 74-96) for the latter group ( P = .024). Reactivation of Infection
Fig. 3. A, Preoperative radiograph of the hip of a 38-year-old woman who had pyogenic hip arthritis at 4 years of age and developed fused hip. B, Radiograph obtained 6 years after THA with a modular stem shows satisfactory fixation of components. However, the patient reports discomfort due to residual leg length difference.
All intraoperative swab and tissue cultures were negative for tuberculous or bacterial infection, and only 3 specimens showed histologic signs of chronic inflammation. The remainder of the specimens showed no presence of any organism and no signs that might suggest acute or chronic inflammation. In 1 of the 3 patients, mildly inflamed hip joint fluid was found in preoperative aspirate, but the patient’s blood count and sedimentation rate were normal, and there was no sign of ongoing inflammation. The patient did not develop recurrent infection after the operation. Infection recurred
612 The Journal of Arthroplasty Vol. 20 No. 5 August 2005 after THA in 1 patient (1.3%) who had had tuberculous hip arthritis; the patient subsequently underwent a Girdlestone hip operation. The interval between active infection and arthroplasty in this patient was 11 years, and the patient’s preoperative sedimentation rate was normal. Radiographic Results One active young patient underwent cup revision because of polyethylene wear and loosening 6 years after the index operation. Three hips showed radiographic evidence of a loose acetabular component. Of those, 2 hips that had required structural autograft and screw fixations for severe acetabular dysplasia developed loosening and subsequently underwent revision. No radiographically loose femoral stem was detected at the final followup examination. Thus, the overall mechanical loosening rate was 5.3%. All of the failed acetabular components were in patients with Crowe group 4 hip dysplasia. Complications One female patient developed sciatic nerve palsy, from which she partially recovered, with mild sensory decrement, by 4 years after surgery. One periprosthetic fracture occurred at the anterior cortex of stem tip in a patient who underwent stem implantation after femoral shortening. The patient was successfully treated with open reduction and internal fixation 2 weeks after the index operation.
Discussion Our study concentrated on prognostic factors for THA performed because of infection sequelae, regardless of type. Because tuberculosis is endemic to Korea, we often found it difficult to clearly distinguish old tuberculous sequelae from pyogenic sequelae of the hip. Both types of infections are frequently encountered in Asian countries and tend to yield similar deformity. Although they may manifest different clinical features at the time of active infection, both types show similar softtissue scarring and destruction of bony structures. Clinical results in this specific group of patients were mainly influenced by age at the time of onset of the infection and extent of preoperative bony deformations (as represented clinically by large leg length difference or fusion of the hip, and radiographically by severe acetabular dysplasia with or without hypoplastic femur). Neither
preoperative flexion contracture nor improved arc of motion after surgery, however, influenced clinical results. Patients’ age at the time of infection onset was related to clinical results. Patients who acquired an infection at 5 years or younger had a lower HHS at final follow-up evaluation, most likely because the infection impeded skeletal growth and the development of anatomic structures, causing a greater degree of deformation. Although we attempted to correct leg length difference in all cases, some discrepancy inadvertently remained in some patients with a relatively larger degree of leg length inequality. When the initial leg length difference was more than 1 in, there was a relatively higher chance that some deformity would remain after surgery and produce a less favorable clinical outcome. Patients with severe acetabular dysplasia (Crowe group 4) had less-than-optimal clinical outcomes, with an average final follow-up HHS of 83.1 (range, 74-90). Four hips (5.3%) had aseptic loosening of the acetabular component, 3 of them subsequently undergoing revision. Of those 3 hips, 2 required structural autograft and screw fixation for severe acetabular dysplasia at the time of primary arthroplasty. Severe anatomic deformation of the acetabulum can be managed with a Charnley cemented cup or with bulk autogenous graft or allograft from the femoral head and insertion of the acetabular component with cement, but the reported results have not been encouraging for either fixation method [13,14]. Anderson and Harris [4], in their intermediateterm follow-up study, reported that stable fixation was achieved with the use of cementless acetabular components fixed with screws in hips with severe acetabular dysplasia, provided that at least 70% of the porous surface was covered by host bone. Because we have had experience with acetabular failure with structural autograft, we implanted 22-mm-head components and small (38-mm, 40-mm, and 42-mm) acetabular cups with medialization to gain better bony coverage in hips with severe acetabular dysplasia. Hypoplastic femur with a small proximal diameter secondary to undergrowth or destruction may limit the size and shape of the stem. Some authors have reported a high rate of revision due to mismatch of stem size with the canal. Kim found a high rate of loosening (17%) with the Mqller CDH femoral prosthesis (Protek) and attributed it to use of small femoral stems in normal femoral canals. Intraoperative fractures have been reported, and it has been suggested that poor initial
Factors Influencing Outcome of THA for Hip Infection Sequelae ! Park et al 613
stability is related to subsequent loosening [5]. Some authors advocate use of subtrochanteric shortening with derotational osteotomy or of a specially designed prosthesis, such as a CDH stem with swan-neck femoral components [1-7]. Such techniques have reportedly improved clinical outcomes in midterm to long-term studies. In our study, with an average follow-up period of 5.8 years, we found no radiologically loose stems. We expect that the loosening rate will remain low in the long term, considering the characteristics of porous-coated implants, which tend to produce persistent initial fixation. We believe that despite the femoral hypoplasia we encountered, the development of special stems, such as a modular design with various sizes, and current surgical techniques were associated with our improved clinical results. In our investigation, recurrence of infection after THA in this group of patients was uncommon; however, functional outcome was compromised by the extent of anatomic deformation caused by the initial infection. Age at the time of infection onset, preoperative leg length difference, presence of hip fusion, presence of severe dysplastic acetabulum as shown on radiographs, and the presence of proximal femoral hypoplasia (reflected in a small femoral isthmus cortical width) may be predictive of functional outcome.
References 1. DiFazio F, Shon WY, Salvati EA, et al. Long-term results of total hip arthroplasty with a cemented custom-designed swan-neck femoral component for congenital dislocation or severe dysplasia: a followup note. J Bone Joint Surg Am 2002;84:204. 2. Xener N, Tfzqn I˙R, AYik M. Femoral shortening and cementless arthroplasty in high congenital dislocation of the hip. J Arthroplasty 2002;17:41.
3. Kim YH. Total hip arthroplasty in adult patients who had childhood infection of the hip. J Bone Joint Surg Am 2003;85:198. 4. Anderson MJ, Harris WH. Total hip arthroplasty with insertion of the acetabular component without cement in hips with total congenital dislocation or marked congenital dysplasia. J Bone Joint Surg Am 1999;81:347. 5. Kim YH. Total arthroplasty of the hip after childhood sepsis. J Bone Joint Surg Br 1991;73:783. 6. Klisic P, Jankovic L. Combined procedure of open reduction and shortening of the femur in treatment of congenital dislocation of the hips in older children. Clin Orthop 1976;119:60. 7. Masonis JL, Patel JV, Miu A, et al. Subtrochanteric shortening and derotational osteotomy in primary total hip arthroplasty for patients with severe hip dysplasia. J Arthroplasty 2003;18 Suppl 1:68. 8. Cherney DL, Amstutz HC. Total hip replacement in the previously septic hip. J Bone Joint Surg Am 1983;65:1256. 9. Hardinge K, Cleary J, Charnley J. Low-friction arthroplasty for healed septic and tuberculous arthritis. J Bone Joint Surg Br 1979;61:144. 10. Jupiter JB, Karchmer AW, Lowell JD, et al. Total hip arthroplasty in the treatment of adult hips with current or quiescent sepsis. J Bone Joint Surg Am 1981;63:194. 11. Crowe JF, Mani VJ, Ranawat CS. Total hip replacement in congenital dislocation and dysplasia of the hip. J Bone Joint Surg 1979;61:15. 12. Harris WH. Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. J Bone Joint Surg Am 1969; 51:737. 13. Gerber SD, Harris WH. Femoral head autografting to augment acetabular deficiency in patients requiring total hip replacement: a minimum five-year and an average seven-year follow-up study. J Bone Joint Surg Am 1986;68:1241. 14. Numair J, Joshi AB, Murphy KC, et al. Total hip arthroplasty for congenital dysplasia or dislocation of the hip: survivorship analysis and long-term results. J Bone Joint Surg Am 1997;79:1352.