- Email: [email protected]
Aspiration as a guide to sepsis in revision total hip arthroplasty
The Journal of Arthroplasty Vol. 11 No. 5 1996
A s p i r a t i o n as a G u i d e to Sepsis in R e v i s i o n Total Hip A r t h r o p l a s t y Thomas
K. Fehring,
MD,* and Bruce Cohen,
MDt
Abstract: One hundred sixty-five patients underwent 171 preoperative aspiration arthrograms to evaluate a painful total hip arthroplasty. Intraoperative cultures and histologic specimens were obtained in all cases. Of the 166 aspirations where fluid was obtained, there were 140 true negative, 5 true positive, 18 false positive, and 3 false negative cultures. Sensitivity of hip aspiration to identify periprosthetic sepsis correctly was 50%; specificity was 88%. Hip aspiration with a 50% sensitivity rate lacks the ability to consistently predict those patients with occult periprosthetic sepsis. The routine use of aspiration in evaluation of a painful total hip is probably not indicated. Selective use in patients with a history of wound healing problems, radiographic changes, and elevated laboratory values should be considered. Key words: hip, aspiration, infection, arthroplasty, sepsis.
Infection is an u n c o m m o n sequela of total hip arthroplasty (THA); however, w h e n evaluating a patient with a painful total hip, this diagnosis m u s t be considered. W h e n a patient presents with obvious loosening and no clinical signs of infection, it is important to determine w h e t h e r loosening is related to a septic or aseptic process. Aspiration has b e e n r e c o m m e n d e d as an integral part of this evaluation. The ability to detect accurately those patients with indolent sepsis has significant therapeutic implications. If those patients can be identified before surgery as having loosened f r o m occult sepsis, better overall revision results m i g h t be realized. Clearly, two-stage r e i m p l a n t a t i o n has a greater success rate t h a n p r i m a r y - e x c h a n g e arthroplasty [1-5]. P r o p o n e n t s of routine aspiration point to these facts in r e c o m m e n d i n g this p r o c e d u r e prior to revision THA. The usefulness of a diagnostic tool such as aspiration is d e t e r m i n e d by its sensitivity or ability to detect an infection w h e n present. We b e c a m e con-
cerned a b o u t the sensitivity of this test in our revision practice after a n u m b e r of false positive aspirations w e r e e n c o u n t e r e d . We therefore sought to review our series of consecutive revisions to determ i n e w h e t h e r hip aspirations should be routinely r e c o m m e n d e d to our patients w h o present w i t h painful total hips.
Materials and Methods B e t w e e n 1989 and 1993, 165 patients u n d e r w e n t preoperative aspiration arthrograms to evaluate a painful THA. All aspirations w e r e p e r f o r m e d u n d e r sterile conditions in a radiology suite by one of three musculoskeletal radiologists. Intraarticular p l a c e m e n t was confirmed by contrast arthrography. Aspirated fluid was sent for cell count, G r a m stain, and aerobic and anaerobic cultures. All 165 patients u n d e r w e n t subsequent revision THA. At the time of surgery, each patient h a d aerobic, anaerobic, acidfast bacillus, and fungal cultures t a k e n during opening of the joint pseudocapsule. Two sets of fluid cultures were t a k e n as was a tissue culture f r o m the joint. Each aerobic and anaerobic swab culture was plated on trypticase soy agar, M a c C o n k e y II agar, chocolate agar, anaerobic blood agar, anaerobic
From the *Charlotte Orthopedic Specialists, t~A., and /-Carolinas Medical Center, Charlotte, North Carolina. Reprint requests: Thomas K. Fehring, MD, Charlotte Orthopedic Specialists, P.A., 2600 East Seventh Street, Charlotte, NC 28204.
543
544
The Journal of Arthroplasty Vol. 11 No. 5 August 1996
blood agar with kanamycin and vancomycin, and thioglycollate broth. The anaerobic and thioglycollate broth specimens were held for 5 days; the remainder were finalized at 3 days. Aerobic and anaerobic tissue cultures were plated on similar media with the addition of a phenylethyl alcohol agar. Tissue cultures for acidfast bacilli were incubated 8 weeks in Lowenstein-Jensen medium and in a Bactec blood culture bottle (Becton Dickinson, Sparks, MD). Fungal tissue cultures were incubated 4 weeks in brain-heart infusion agar and inhibitory mold agar. In 130 patients, frozen sections were taken from the joint pseudocapsule and from the membrane of a loose component. A complete histologic analysis was also performed, which included permanent sections prepared from the tissue submitted for frozen section as well as widely sampled sites from additional tissue resected. All patients with positive cultures were evaluated by an infectious disease consultant. All histologic sections were interpreted by a single musculoskeletal pathologist. Histologic sections were considered positive for infection if there was evidence of acute inflammation characterized by the presence of polymorphonuclear leukocytes. Most negative sections had a histologic picture of dense fibrosis, focal necrosis, or a histiocytic response to embedded foreign material. Positive cultures corroborated with histologic sections, clinical examination, and appropriate laboratory tests were considered pathogens. The specific definition of true infection is difficult to determine. It is important to recognize that no single test is adequate to diagnose infection. Multiple parameters must be considered. Carlsson et al. described their diagnostic criteria for infection to include culture results and clinical and laboratory features, without including histologic examination [6]. In a previous work regarding frozen histologic section, we found frozen section to have a low sensitivity in detecting occult sepsis w he n present; however, the sensitivity and specificity of complete histologic evaluation were high [7]. We therefore chose to include the complete histologic evaluation as an integral part of our determination of whether sepsis was present. Preoperative evaluation of these patients included a detailed history stressing perioperative wound healing and a physical examination with particular attention to swelling, warmth, drainage, fever, or other constitutional symptoms. Anteroposterior and lateral radiographs including the entire implant were routinely obtained. A complete blood cell count, erythrocyte sedimentation rate, and C-reactive protein were also routinely obtained.
Revision THA was performed in an operating suite with horizontal laminar flow. Intraoperative antibiotics were withheld until appropriate cultures were taken and were continued after surgery until final cultures and histology were completed. A retrospective review of 165 patients was undertaken to determine the sensitivity and specificity of hip aspiration as a diagnostic tool in detecting occult periprosthetic infection during revision arthroplasty. The relationships between hip aspiration cultures, intraoperative cultures, and histologic sections were also evaluated. A total of 171 aspirations were performed in 165 patients. Five patients had dry taps, leaving a data set of i66 aspirations. Subsequent intraoperative cultures and histologic analyses were negative for active infection in these five patients. Six patients had repeat aspirations performed. These were performed because of either a high clinical suspicion and a negative initial aspiration or a broth only result in a patient with no risk factors. Four of these six patients' repeat aspirations were negative. The remaining two patients had a positive culture on repeat aspiration. Both of these patients had true infections. Intraoperative cultures that were positive in the liquid medium only were not considered diagnostic unless final pathologic examination demonstrated a significant amount of inflammatory reaction. As noted above, positive cultures that corroborated with histologic sections, clinical examination, and appropriate laboratory tests were considered pathogens. Four patients had true infection based on these strict criteria. ~vo other patients had negative intraoperative cultures yet had strong histologic evidence of infection. Each of these two patients also had a previous history of periprosthetic sepsis, and therefore, clinical suspicion was high. These patients were considered to have true infections by our infectious disease consultants and were treated accordingly.
Results Of the 166 aspirates where fluid was obtained, there were 140 true negative cultures, 3 true positive cultures (2 patients reaspirated), 18 false positive cultures, and 3 false negative cultures. Sensitivity (defined as true positives divided by true positives plus false negatives) was 50%. The standard error for the sensitivity of hip aspiration was _+ 20% with a 95% confidence interval of 12 to 92%. This large confidence interval is a reflection of the low prevalence of actual infections in our series. Specificity (true negatives divided by true
Aspirationas Guide to Sepsis in RevisionTHA • negatives plus false positives) was 88%. The standard error for the specificity of hip aspiration was _+ 2% with a 95% confidence interval of 82 to 93%. The accuracy was 87% with a standard error of _+ 3% and a 95% confidence interval of 8I to 92%. The organisms cultured in the false positive aspirate group included predominantly growth in liquid m e d i u m only. Staphylococcus epidermidis dominated these cultures, being present in 12 cases (Table 1). Intraoperative cultures were f o u n d to be false positive in 19 patients. The majority of these were culture positive in liquid m e d i u m only and considered nondiagnostic because their clinical history and pathologic examination were not consistent with infection. There were 2 false negative cultures, 140 true negative cultures, 20 false positive cultures, and 4 true positive cultures in the intraoperative culture results. The sensitivity of intraoperative cultures was 67% with a standard error of _+ 21% and a 95% confidence interval of 22 to 96%. Once again, the large standard error and confidence interval are a reflection of the low prevalence of actual infections in this series. The specificity was 88% with a standard error of _+ 3% and a confidence interval of 82 to 93%. The accuracy was 87% with a standard error of + 3% and a confidence interval of 81 to 92%. False positive organisms included similar pathogens as n o t e d in the false positive aspiration data (Table 2). Frozen histologic sections were obtained in I30 patients. These included 114 true negative results, 2 true positive results, 12 false positive results, and 2 false negative results. This resulted in a sensitivity of 50 _+ 29% with a 95% confidence interval of 7 to 93%. The specificity was 90 _ 3% with a confidence interval of 84 to 95%. The accuracy was 89 _+3% with a confidence interval of 82 to 94%.
Discussion The ability to detect accurately occult periprosthetic infection w h e n present is important in the evaluation of a patient presenting with a painful THA. Unfortunately, no single m e t h o d is able to
Table 1. Organisms in False Positive Aspirations
Staphylococcus epidermidis Proprionibacterium Candida
Group D streptococci
Total
Broth Only
12 4 2
9 4 2
1
1
Fehring and Cohen
545
Table 2. Organisms in False Positive Intraoperative Cultures
Staphylococcus epiderrnidis Proprionibacterium Aspergillus Bacillus species y-Hemolytic streptococci Corynebacterium Enterobacter Group D streptococci
Total
Broth Only
Ii 2 1 1 2 1 1 I
8 1 1 1 1 1 1 1
reproducibly distinguish between septic and aseptic loosening. A careful history can provide the first indication of indolent sepsis. A history of difficult w o u n d healing, prolonged drainage, rest pain, or lack of a pain-free interval following surgery should raise suspicion in the mind of the examiner. Physical findings, in the case of indolent sepsis, are usually nondescript. W h e n cellulitis, induration, or active drainage are present, the diagnosis is not dilficult. Serial laboratory tests are important in the evaluation of any revision patient. The white blood cell count or the erythrocyte sedimentation rate m a y be elevated in septic cases. The erythrocyte sedim e n t a t i o n rate was greater t h a n 30 m m / h in 75% of 65 patients with infected total hips reported by McDonald et al. [8]. Its accuracy in determining occult sepsis, however, was questioned by Canner et al., w h o n o t e d that in 46% of 52 patients with active sepsis the sedimentation rate was less t h a n 30 [3]. C-reactive protein m a y be the most sensitive laboratory test, as it should r e t u r n to n o r m a l within 3 weeks of surgery in aseptic cases [9]. Plain radiographs are particularly helpful if carefully scrutinized for periosteal elevation, endosteal scalloping, or areas of osteolysis. Early radiographic deterioration in an otherwise well-performed arthroplasty should alert one to the possibility of a septic process. Aspiration of the joint can provide direct bacteriologic evidence of occult periprosthetic sepsis. This knowledge can guide the surgeon in outlining an appropriate t r e a t m e n t plan for the patient with a painful total hip. In theory, if the cultures are negative, the surgeon m a y proceed with the reassurance that she or he is dealing with aseptic loosening. II the cultures are positive, the surgeon can decide w h e t h e r a one- or two-stage reimplantation is appropriate based o n the offending organism and clinical situation. The ability of hip aspiration to adequately provide this information is debated in the literature. Its usefulness as a diagnostic test has led some
546
The Journal of Arthroplasty Vol. 11 No. 5 August 1996
authors to recommend this study as an essential part of the preoperative evaluation [10] and others to dismiss it as unreliable [11]. Phillips et al. noted 83% correct results in their evaluation of 141 joint aspirations [10]. In that study, they observed a false positive rate of 16%, yet still considered preoperative aspiration an essential part of the evaluation of patients who experienced pain after THA. Bucholz et al., in comparing results of fluid aspirated with cultures obtained at surgery, noted that the aspirate result was correct in 73% of cases [2]. McDonald et al. reported similar results, noting aspirations were positive in 60% of infected THAs [8]. Weber and Lautenback found aspirations able to correctly predict the offending organism in only 55 % of chronically infected THAs [12]. Wroblewski noted that aspirations were unreliable in his study dealing with one-stage revisions of infected total hips and, therefore, did not recommend them as a staging procedure [11]. Barrack and Harris also questioned the value of aspiration prior to revision THA in their review of 270 patients [1]. They noted a 13% false positive rate in their study. Only two true positive findings were noted, and less than 10% of the positive aspirates in their study were considered pathogens. Based on their findings, they believed aspiration should be performed selectively rather than routinely. They also believed aspiration should be performed only if the history suggested infection or the radiographs demonstrated lysis or periostitis. Every infection in their series would have been identified had these specific indications been followed. The number of false positive aspirations presented in our series also casts doubt on the ability of this diagnostic test to provide accurate preoperative information in evaluating a painful total hip. We noted an 11% false positive rate, and less than 25% of the 23 positive aspirates in our study were considered pathogens.
Table
Patient No. 1 2 3 4 5 6
C-Reactive Protein NA 5.3 < 1 <1 2.7 8
NA, not applicable.
3.
Erythrocyte Sedimentation Rate 52 112 42 21 35 91
Six Patients Wound Healing Problems + + + +
With
We were equally concerned about the three false negative aspirations noted from our cohort of six true infections. In those patients with a negative initial aspirate and strong clinical or radiographic suspicion of occult sepsis, a repeat aspiration may have been in order. As the majority of false positive aspirations were broth only, this fact alone did not prevent reimplantation. Without corroborating clinical, radiographic, or laboratory evidence of infection, reimplantation usually proceeded with the assumption that we were dealing with a contaminate on aspirate. The false negative results were more troublesome in that the realization of true infection came after surgery, following evaluation of intraoperative cultures and final pathology. Having unknowingly performed a primary exchange makes subsequent management more complicated. Using the criteria suggested by Barrack and Harris [1] in performing selective aspirations, each of our patients with true infections would have been aspirated. In our series, each patient with a true infection had some clinical, laboratory, or radiographic evidence that would have prompted us to order an aspiration before surgery (Table 3). Four of our six patients with true infections had wound healing problems. Five of our six patients had elevated C-reactive proteins or sedimentation rates, and one had radiographic evidence of osteolysis.
Conclusion The ability of a diagnostic test to predict accurately the presence of a disease process is determined by its sensitivity. Aspiration, with a sensitivity of only 50% in our study, may be lacking in its ability to consistently predict those patients with occult periprosthetic sepsis; however, it is important to realize in evaluating our statistics that there is a low prevalence of actual infection in this cohort of patients. Therefore, calculations of sensi-
Infected
Revision
Radiographic Changes + -
Arthroplasty
Aspiration + + +
Intraoperative Culture + + + +
Final Pathology + + + + + +
Aspiration as Guide to Sepsis in RevisionTHA
tivity are susceptible to bias a n d m a y lack statistical power. On the other hand, hip aspiration can be considered to be fairly specific in its ability to determ i n e the absence of infection. Therefore, should one routinely use a diagnostic test w h o s e sensitivity is in question a n d w h o s e specificity is fairly high? Or p u t a n o t h e r way, should we aspirate 165 hips to possibly diagnose 6 true infections, r e m e m bering that 3 of these true infections h a d negative aspirations a n d there was a high rate of false positives in o u r series? We agree w i t h previous investigators that aspiration in revision TtlA should be p e r f o r m e d selectively r a t h e r t h a n routinely [1]. All of o u r patients w i t h true infections w o u l d h a v e b e e n aspirated using the criteria of clinical, laboratory, or radiographic suspicion. A detailed history, s e r u m laboratory tests, and careful radiographic analysis should delineate those patients requiring aspirations. If clinical suspicion persists despite a negative initial aspirate, repeat aspiration should be considered.
References 1. Barrack RL, Harris WH: The value of aspiration of the hip before total hip arthroplasty. J Bone Joint Surg 75A:66, 1993 2. Bucholz HW, Elson RA, Engelbrecht E et al: Management of Deep Infection of Total Hip Replacement. J Bone Joint Surg 63B:342, 1981
•
Fehring and Cohen
547
3. Canner GC, Steinberg ME, Heppenstall RB, Baldersten R: Infected hip after total hip arthroplasty. J Bone Joint Surg 66A:1393, 1984 4. Dupont JA: Significance of operative cultures in total hip arthroplasty. Clin Orthop 211:122, 1986 5. Harwood DA, Robins SG, Zawadsky JP, Bullek D: Predictive values of intraoperative and postoperative cultures in patients undergoing total arthroplasty. Orthop Trans 13:60, 1989 6. Carlsson AS, Josefsson G, Lundberg G: Revision with gentamicin-impregnated cement for deep infections in total hip arthroplasties. J Bone Joint Surg 60A:1059, 1978 7. Fehring TK, McAlister JA Jr: Frozen histologic section as a guide to sepsis in revision joint arthroplasty. Clin Orthop 301:229, 1994 8. McDonald DJ, Fitzgerald RH Jr, Ilstrup DM: Twostage reconstruction of a total hip arthroplasty because of infection. J Bone Joint Surg 71A:828, 1989 9. Sansen L: Erythrocyte sedimentation rate following exchange of infected total hips. Acta Orthop Scand 59:148, 1988 10. Phillips W, Kattapuram SV: Efficacy of preoperative hip aspiration performed in the radiology department. Clin Orthop 179:141, 1983 11. Wroblewski BM: One-stage revision of infected cemented total hip arthroplasty. Clin Orthop 211:103, 1986 12. Weber FA, Lautenback LEG: Revision of infected total hip arthroplasty. Clin Orthop 211:108, 1986
A s p i r a t i o n as a G u i d e to Sepsis in R e v i s i o n Total Hip A r t h r o p l a s t y Thomas
K. Fehring,
MD,* and Bruce Cohen,
MDt
Abstract: One hundred sixty-five patients underwent 171 preoperative aspiration arthrograms to evaluate a painful total hip arthroplasty. Intraoperative cultures and histologic specimens were obtained in all cases. Of the 166 aspirations where fluid was obtained, there were 140 true negative, 5 true positive, 18 false positive, and 3 false negative cultures. Sensitivity of hip aspiration to identify periprosthetic sepsis correctly was 50%; specificity was 88%. Hip aspiration with a 50% sensitivity rate lacks the ability to consistently predict those patients with occult periprosthetic sepsis. The routine use of aspiration in evaluation of a painful total hip is probably not indicated. Selective use in patients with a history of wound healing problems, radiographic changes, and elevated laboratory values should be considered. Key words: hip, aspiration, infection, arthroplasty, sepsis.
Infection is an u n c o m m o n sequela of total hip arthroplasty (THA); however, w h e n evaluating a patient with a painful total hip, this diagnosis m u s t be considered. W h e n a patient presents with obvious loosening and no clinical signs of infection, it is important to determine w h e t h e r loosening is related to a septic or aseptic process. Aspiration has b e e n r e c o m m e n d e d as an integral part of this evaluation. The ability to detect accurately those patients with indolent sepsis has significant therapeutic implications. If those patients can be identified before surgery as having loosened f r o m occult sepsis, better overall revision results m i g h t be realized. Clearly, two-stage r e i m p l a n t a t i o n has a greater success rate t h a n p r i m a r y - e x c h a n g e arthroplasty [1-5]. P r o p o n e n t s of routine aspiration point to these facts in r e c o m m e n d i n g this p r o c e d u r e prior to revision THA. The usefulness of a diagnostic tool such as aspiration is d e t e r m i n e d by its sensitivity or ability to detect an infection w h e n present. We b e c a m e con-
cerned a b o u t the sensitivity of this test in our revision practice after a n u m b e r of false positive aspirations w e r e e n c o u n t e r e d . We therefore sought to review our series of consecutive revisions to determ i n e w h e t h e r hip aspirations should be routinely r e c o m m e n d e d to our patients w h o present w i t h painful total hips.
Materials and Methods B e t w e e n 1989 and 1993, 165 patients u n d e r w e n t preoperative aspiration arthrograms to evaluate a painful THA. All aspirations w e r e p e r f o r m e d u n d e r sterile conditions in a radiology suite by one of three musculoskeletal radiologists. Intraarticular p l a c e m e n t was confirmed by contrast arthrography. Aspirated fluid was sent for cell count, G r a m stain, and aerobic and anaerobic cultures. All 165 patients u n d e r w e n t subsequent revision THA. At the time of surgery, each patient h a d aerobic, anaerobic, acidfast bacillus, and fungal cultures t a k e n during opening of the joint pseudocapsule. Two sets of fluid cultures were t a k e n as was a tissue culture f r o m the joint. Each aerobic and anaerobic swab culture was plated on trypticase soy agar, M a c C o n k e y II agar, chocolate agar, anaerobic blood agar, anaerobic
From the *Charlotte Orthopedic Specialists, t~A., and /-Carolinas Medical Center, Charlotte, North Carolina. Reprint requests: Thomas K. Fehring, MD, Charlotte Orthopedic Specialists, P.A., 2600 East Seventh Street, Charlotte, NC 28204.
543
544
The Journal of Arthroplasty Vol. 11 No. 5 August 1996
blood agar with kanamycin and vancomycin, and thioglycollate broth. The anaerobic and thioglycollate broth specimens were held for 5 days; the remainder were finalized at 3 days. Aerobic and anaerobic tissue cultures were plated on similar media with the addition of a phenylethyl alcohol agar. Tissue cultures for acidfast bacilli were incubated 8 weeks in Lowenstein-Jensen medium and in a Bactec blood culture bottle (Becton Dickinson, Sparks, MD). Fungal tissue cultures were incubated 4 weeks in brain-heart infusion agar and inhibitory mold agar. In 130 patients, frozen sections were taken from the joint pseudocapsule and from the membrane of a loose component. A complete histologic analysis was also performed, which included permanent sections prepared from the tissue submitted for frozen section as well as widely sampled sites from additional tissue resected. All patients with positive cultures were evaluated by an infectious disease consultant. All histologic sections were interpreted by a single musculoskeletal pathologist. Histologic sections were considered positive for infection if there was evidence of acute inflammation characterized by the presence of polymorphonuclear leukocytes. Most negative sections had a histologic picture of dense fibrosis, focal necrosis, or a histiocytic response to embedded foreign material. Positive cultures corroborated with histologic sections, clinical examination, and appropriate laboratory tests were considered pathogens. The specific definition of true infection is difficult to determine. It is important to recognize that no single test is adequate to diagnose infection. Multiple parameters must be considered. Carlsson et al. described their diagnostic criteria for infection to include culture results and clinical and laboratory features, without including histologic examination [6]. In a previous work regarding frozen histologic section, we found frozen section to have a low sensitivity in detecting occult sepsis w he n present; however, the sensitivity and specificity of complete histologic evaluation were high [7]. We therefore chose to include the complete histologic evaluation as an integral part of our determination of whether sepsis was present. Preoperative evaluation of these patients included a detailed history stressing perioperative wound healing and a physical examination with particular attention to swelling, warmth, drainage, fever, or other constitutional symptoms. Anteroposterior and lateral radiographs including the entire implant were routinely obtained. A complete blood cell count, erythrocyte sedimentation rate, and C-reactive protein were also routinely obtained.
Revision THA was performed in an operating suite with horizontal laminar flow. Intraoperative antibiotics were withheld until appropriate cultures were taken and were continued after surgery until final cultures and histology were completed. A retrospective review of 165 patients was undertaken to determine the sensitivity and specificity of hip aspiration as a diagnostic tool in detecting occult periprosthetic infection during revision arthroplasty. The relationships between hip aspiration cultures, intraoperative cultures, and histologic sections were also evaluated. A total of 171 aspirations were performed in 165 patients. Five patients had dry taps, leaving a data set of i66 aspirations. Subsequent intraoperative cultures and histologic analyses were negative for active infection in these five patients. Six patients had repeat aspirations performed. These were performed because of either a high clinical suspicion and a negative initial aspiration or a broth only result in a patient with no risk factors. Four of these six patients' repeat aspirations were negative. The remaining two patients had a positive culture on repeat aspiration. Both of these patients had true infections. Intraoperative cultures that were positive in the liquid medium only were not considered diagnostic unless final pathologic examination demonstrated a significant amount of inflammatory reaction. As noted above, positive cultures that corroborated with histologic sections, clinical examination, and appropriate laboratory tests were considered pathogens. Four patients had true infection based on these strict criteria. ~vo other patients had negative intraoperative cultures yet had strong histologic evidence of infection. Each of these two patients also had a previous history of periprosthetic sepsis, and therefore, clinical suspicion was high. These patients were considered to have true infections by our infectious disease consultants and were treated accordingly.
Results Of the 166 aspirates where fluid was obtained, there were 140 true negative cultures, 3 true positive cultures (2 patients reaspirated), 18 false positive cultures, and 3 false negative cultures. Sensitivity (defined as true positives divided by true positives plus false negatives) was 50%. The standard error for the sensitivity of hip aspiration was _+ 20% with a 95% confidence interval of 12 to 92%. This large confidence interval is a reflection of the low prevalence of actual infections in our series. Specificity (true negatives divided by true
Aspirationas Guide to Sepsis in RevisionTHA • negatives plus false positives) was 88%. The standard error for the specificity of hip aspiration was _+ 2% with a 95% confidence interval of 82 to 93%. The accuracy was 87% with a standard error of _+ 3% and a 95% confidence interval of 8I to 92%. The organisms cultured in the false positive aspirate group included predominantly growth in liquid m e d i u m only. Staphylococcus epidermidis dominated these cultures, being present in 12 cases (Table 1). Intraoperative cultures were f o u n d to be false positive in 19 patients. The majority of these were culture positive in liquid m e d i u m only and considered nondiagnostic because their clinical history and pathologic examination were not consistent with infection. There were 2 false negative cultures, 140 true negative cultures, 20 false positive cultures, and 4 true positive cultures in the intraoperative culture results. The sensitivity of intraoperative cultures was 67% with a standard error of _+ 21% and a 95% confidence interval of 22 to 96%. Once again, the large standard error and confidence interval are a reflection of the low prevalence of actual infections in this series. The specificity was 88% with a standard error of _+ 3% and a confidence interval of 82 to 93%. The accuracy was 87% with a standard error of + 3% and a confidence interval of 81 to 92%. False positive organisms included similar pathogens as n o t e d in the false positive aspiration data (Table 2). Frozen histologic sections were obtained in I30 patients. These included 114 true negative results, 2 true positive results, 12 false positive results, and 2 false negative results. This resulted in a sensitivity of 50 _+ 29% with a 95% confidence interval of 7 to 93%. The specificity was 90 _ 3% with a confidence interval of 84 to 95%. The accuracy was 89 _+3% with a confidence interval of 82 to 94%.
Discussion The ability to detect accurately occult periprosthetic infection w h e n present is important in the evaluation of a patient presenting with a painful THA. Unfortunately, no single m e t h o d is able to
Table 1. Organisms in False Positive Aspirations
Staphylococcus epidermidis Proprionibacterium Candida
Group D streptococci
Total
Broth Only
12 4 2
9 4 2
1
1
Fehring and Cohen
545
Table 2. Organisms in False Positive Intraoperative Cultures
Staphylococcus epiderrnidis Proprionibacterium Aspergillus Bacillus species y-Hemolytic streptococci Corynebacterium Enterobacter Group D streptococci
Total
Broth Only
Ii 2 1 1 2 1 1 I
8 1 1 1 1 1 1 1
reproducibly distinguish between septic and aseptic loosening. A careful history can provide the first indication of indolent sepsis. A history of difficult w o u n d healing, prolonged drainage, rest pain, or lack of a pain-free interval following surgery should raise suspicion in the mind of the examiner. Physical findings, in the case of indolent sepsis, are usually nondescript. W h e n cellulitis, induration, or active drainage are present, the diagnosis is not dilficult. Serial laboratory tests are important in the evaluation of any revision patient. The white blood cell count or the erythrocyte sedimentation rate m a y be elevated in septic cases. The erythrocyte sedim e n t a t i o n rate was greater t h a n 30 m m / h in 75% of 65 patients with infected total hips reported by McDonald et al. [8]. Its accuracy in determining occult sepsis, however, was questioned by Canner et al., w h o n o t e d that in 46% of 52 patients with active sepsis the sedimentation rate was less t h a n 30 [3]. C-reactive protein m a y be the most sensitive laboratory test, as it should r e t u r n to n o r m a l within 3 weeks of surgery in aseptic cases [9]. Plain radiographs are particularly helpful if carefully scrutinized for periosteal elevation, endosteal scalloping, or areas of osteolysis. Early radiographic deterioration in an otherwise well-performed arthroplasty should alert one to the possibility of a septic process. Aspiration of the joint can provide direct bacteriologic evidence of occult periprosthetic sepsis. This knowledge can guide the surgeon in outlining an appropriate t r e a t m e n t plan for the patient with a painful total hip. In theory, if the cultures are negative, the surgeon m a y proceed with the reassurance that she or he is dealing with aseptic loosening. II the cultures are positive, the surgeon can decide w h e t h e r a one- or two-stage reimplantation is appropriate based o n the offending organism and clinical situation. The ability of hip aspiration to adequately provide this information is debated in the literature. Its usefulness as a diagnostic test has led some
546
The Journal of Arthroplasty Vol. 11 No. 5 August 1996
authors to recommend this study as an essential part of the preoperative evaluation [10] and others to dismiss it as unreliable [11]. Phillips et al. noted 83% correct results in their evaluation of 141 joint aspirations [10]. In that study, they observed a false positive rate of 16%, yet still considered preoperative aspiration an essential part of the evaluation of patients who experienced pain after THA. Bucholz et al., in comparing results of fluid aspirated with cultures obtained at surgery, noted that the aspirate result was correct in 73% of cases [2]. McDonald et al. reported similar results, noting aspirations were positive in 60% of infected THAs [8]. Weber and Lautenback found aspirations able to correctly predict the offending organism in only 55 % of chronically infected THAs [12]. Wroblewski noted that aspirations were unreliable in his study dealing with one-stage revisions of infected total hips and, therefore, did not recommend them as a staging procedure [11]. Barrack and Harris also questioned the value of aspiration prior to revision THA in their review of 270 patients [1]. They noted a 13% false positive rate in their study. Only two true positive findings were noted, and less than 10% of the positive aspirates in their study were considered pathogens. Based on their findings, they believed aspiration should be performed selectively rather than routinely. They also believed aspiration should be performed only if the history suggested infection or the radiographs demonstrated lysis or periostitis. Every infection in their series would have been identified had these specific indications been followed. The number of false positive aspirations presented in our series also casts doubt on the ability of this diagnostic test to provide accurate preoperative information in evaluating a painful total hip. We noted an 11% false positive rate, and less than 25% of the 23 positive aspirates in our study were considered pathogens.
Table
Patient No. 1 2 3 4 5 6
C-Reactive Protein NA 5.3 < 1 <1 2.7 8
NA, not applicable.
3.
Erythrocyte Sedimentation Rate 52 112 42 21 35 91
Six Patients Wound Healing Problems + + + +
With
We were equally concerned about the three false negative aspirations noted from our cohort of six true infections. In those patients with a negative initial aspirate and strong clinical or radiographic suspicion of occult sepsis, a repeat aspiration may have been in order. As the majority of false positive aspirations were broth only, this fact alone did not prevent reimplantation. Without corroborating clinical, radiographic, or laboratory evidence of infection, reimplantation usually proceeded with the assumption that we were dealing with a contaminate on aspirate. The false negative results were more troublesome in that the realization of true infection came after surgery, following evaluation of intraoperative cultures and final pathology. Having unknowingly performed a primary exchange makes subsequent management more complicated. Using the criteria suggested by Barrack and Harris [1] in performing selective aspirations, each of our patients with true infections would have been aspirated. In our series, each patient with a true infection had some clinical, laboratory, or radiographic evidence that would have prompted us to order an aspiration before surgery (Table 3). Four of our six patients with true infections had wound healing problems. Five of our six patients had elevated C-reactive proteins or sedimentation rates, and one had radiographic evidence of osteolysis.
Conclusion The ability of a diagnostic test to predict accurately the presence of a disease process is determined by its sensitivity. Aspiration, with a sensitivity of only 50% in our study, may be lacking in its ability to consistently predict those patients with occult periprosthetic sepsis; however, it is important to realize in evaluating our statistics that there is a low prevalence of actual infection in this cohort of patients. Therefore, calculations of sensi-
Infected
Revision
Radiographic Changes + -
Arthroplasty
Aspiration + + +
Intraoperative Culture + + + +
Final Pathology + + + + + +
Aspiration as Guide to Sepsis in RevisionTHA
tivity are susceptible to bias a n d m a y lack statistical power. On the other hand, hip aspiration can be considered to be fairly specific in its ability to determ i n e the absence of infection. Therefore, should one routinely use a diagnostic test w h o s e sensitivity is in question a n d w h o s e specificity is fairly high? Or p u t a n o t h e r way, should we aspirate 165 hips to possibly diagnose 6 true infections, r e m e m bering that 3 of these true infections h a d negative aspirations a n d there was a high rate of false positives in o u r series? We agree w i t h previous investigators that aspiration in revision TtlA should be p e r f o r m e d selectively r a t h e r t h a n routinely [1]. All of o u r patients w i t h true infections w o u l d h a v e b e e n aspirated using the criteria of clinical, laboratory, or radiographic suspicion. A detailed history, s e r u m laboratory tests, and careful radiographic analysis should delineate those patients requiring aspirations. If clinical suspicion persists despite a negative initial aspirate, repeat aspiration should be considered.
References 1. Barrack RL, Harris WH: The value of aspiration of the hip before total hip arthroplasty. J Bone Joint Surg 75A:66, 1993 2. Bucholz HW, Elson RA, Engelbrecht E et al: Management of Deep Infection of Total Hip Replacement. J Bone Joint Surg 63B:342, 1981
•
Fehring and Cohen
547
3. Canner GC, Steinberg ME, Heppenstall RB, Baldersten R: Infected hip after total hip arthroplasty. J Bone Joint Surg 66A:1393, 1984 4. Dupont JA: Significance of operative cultures in total hip arthroplasty. Clin Orthop 211:122, 1986 5. Harwood DA, Robins SG, Zawadsky JP, Bullek D: Predictive values of intraoperative and postoperative cultures in patients undergoing total arthroplasty. Orthop Trans 13:60, 1989 6. Carlsson AS, Josefsson G, Lundberg G: Revision with gentamicin-impregnated cement for deep infections in total hip arthroplasties. J Bone Joint Surg 60A:1059, 1978 7. Fehring TK, McAlister JA Jr: Frozen histologic section as a guide to sepsis in revision joint arthroplasty. Clin Orthop 301:229, 1994 8. McDonald DJ, Fitzgerald RH Jr, Ilstrup DM: Twostage reconstruction of a total hip arthroplasty because of infection. J Bone Joint Surg 71A:828, 1989 9. Sansen L: Erythrocyte sedimentation rate following exchange of infected total hips. Acta Orthop Scand 59:148, 1988 10. Phillips W, Kattapuram SV: Efficacy of preoperative hip aspiration performed in the radiology department. Clin Orthop 179:141, 1983 11. Wroblewski BM: One-stage revision of infected cemented total hip arthroplasty. Clin Orthop 211:103, 1986 12. Weber FA, Lautenback LEG: Revision of infected total hip arthroplasty. Clin Orthop 211:108, 1986