Abstract
A prospective study was performed to determine the reliability of analysis of intraoperative frozen sections for the identification of infection during 175 consecutive revision total joint arthroplasties (142 hip and thirty-three knee). The mean interval between the primary and the revision arthroplasty was 7.3 years (range, three months to twenty-three years). To reduce selections bias, tissue was obtained for frozen sections during all revisions in patients who did not have active drainage from the wound or a sinus tract.Of the 175 patients, twenty-three had at least five polymorphonuclear leukocytes per high-power field on analysis of the frozen sections and were considered to have an infection. Of these twenty-three, five had five to nine polymorphonuclear leukocytes per high-power field and eighteen had at least ten polymorphonuclear leukocytes per high-power field. The frozen sections for the remaining 152 patients were considered negative. On the basis of cultures of specimens obtained at the time of the revision operation, nineteen of the 175 patients were considered to have an infection. Of the 152 patients who had negative frozen sections, three were considered to have an infection on the basis of the results of the final cultures. Of the twenty-three patients who had positive frozen sections, sixteen were considered to have an infection on the basis of the results of the final cultures; all sixteen had frozen sections that had demonstrated at least ten polymorphonuclear leukocytes per high-power field.The sensitivity and specificity of the frozen sections were similar regardless of whether an index of five or ten polymorphonuclear leukocytes per high-power field was used. Analysis of the frozen sections had a sensitivity of 84 per cent for both indices, whereas the specificity was 96 per cent when the index was five polymorphonuclear leukocytes and 99 per cent when it was ten polymorphonuclear leukocytes. However, the positive predictive value of the frozen sections increased significantly (p < 0.05), from 70 to 89 per cent, when the index increased from five to ten polymorphonuclear leukocytes per high-power field. The negative predictive value of the frozen sections was 98 per cent for both indices.The current study suggests that it is valuable to obtain tissue for intraoperative frozen sections during revision hip and knee arthroplasty. At least ten polymorphonuclear leukocytes per high-power field was predictive of infection, while five to nine polymorphonuclear leukocytes per high-power field was not necessarily consistent with infection. Less than five polymorphonuclear leukocytes per high-power field reliably indicated the absence of infection.
The decision to proceed with reimplantation at the time of a revision arthroplasty is often confounded by a preliminary workup that failed to distinguish clearly between aseptic and septic loosening. There is currently no universally accepted method for the preoperative or intraoperative determination of infection. While they are of some value, clinical examination, laboratory studies, plain radiographs, nuclear scans, and preoperative aspiration cannot be relied on to demonstrate accurately the presence or absence of infection5,14,18.
Previously reported retrospective studies have suggested that analysis of intraoperative frozen sections is a reliable means of identifying active infection during a revision hip or knee arthroplasty4,6,9,18. The goal of the current study was to determine prospectively the usefulness of intraoperative frozen sections for this purpose and to ascertain the number of polymorphonuclear leukocytes per high-power field that is the best indicator of active infection.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Winner of The American Orthopaedic Association-Zimmer Annual Travel Award for Orthopaedic Residents, June 1995.
‡Departments of Orthopaedic Surgery (J. H. L., P. E. D., and J. D. Z.) and Pathology (P. D. and G. S.), Hospital for Joint Diseases, 301 East 17th Street, New York, N.Y. 10003.
One hundred and seventy-five consecutive patients (142 hips and thirty-three knees) who had a revision total joint arthroplasty between February 1993 and March 1994 were studied prospectively. There were sixty-eight men and 107 women; the mean age was sixty-six years (range, thirty-six to eighty-eight years). The mean interval between the primary and the revision arthroplasty was 7.3 years (range, three months to twenty-three years).
Selection bias was reduced by obtaining tissue for frozen sections during all hip and knee revision arthroplasties, regardless of the results of the preoperative workup for infection, except if there was active drainage from the wound or a sinus tract. On the basis of this criterion, three patients (one hip and two knees) were excluded from the study. All sections were obtained at the time of the initial revision operation; frozen sections obtained during any subsequent procedures were not included in the analysis of the data.
A protocol was established to ensure consistency in tissue sampling. Specimens were routinely taken from the pseudocapsule, areas considered suspicious by the surgeon, and tissue that was identified at the various host-implant interfaces. Frozen sections of multiple samples of tissue were studied with a standard staining technique with use of hematoxylin and eosin. As described by Feldman et al., the sections were analyzed according to several criteria to minimize sampling error: (1) granulation tissue was preferentially analyzed, although at times only dense fibrous or fibrin-rich tissue had been obtained and was therefore evaluated; (2) at least two samples of tissue were used; (3) the five most cellular fields (determined on the basis of the number of polymorphonuclear leukocytes) were analyzed; (4) cell counts were performed under high-power magnification of forty times; and (5) only polymorphonuclear leukocytes, identified within tissue rather than fibrin, with well defined cytoplasmic borders were included in the count. (Cells with isolated nuclear fragmentation were excluded as they could not be definitely categorized as polymorphonuclear leukocytes.)
In order to avoid additional inconsistencies in the tissue-sampling, care was taken to obtain swabs for culture from the same tissues that were to be analyzed histologically.
After the frozen sections had been analyzed, permanent sections were prepared by embedding the tissue specimen in paraffin and slicing it approximately every three millimeters into sections that were four to five micrometers thick. The results of the histological analysis of the permanent sections were classified as negative or positive depending on whether there were less than five or at least five polymorphonuclear leukocytes per high-power field.
The frozen section was also considered negative if there were less than five polymorphonuclear leukocytes per high-power field and as positive if there were at least five polymorphonuclear leukocytes per high-power field (Fig. 1). Each positive frozen section was categorized further as having five to nine or at least ten polymorphonuclear leukocytes per high-power field. The results of analysis of the intraoperative frozen sections were used by the surgeon to guide the decision as to whether to proceed with primary exchange of the implant or to perform a resection arthroplasty and a staged reimplantation. For three patients, the results of culture of a preoperative aspirate, which did not correspond with the negative intraoperative frozen sections, determined treatment. A fourth patient had equivocal frozen sections (five to nine polymorphonuclear leukocytes per high-power field) despite a preoperative workup that had suggested the absence of infection; this patient was managed on the basis of the gross appearance of the tissue and the results of the comprehensive preoperative evaluation.
Antibiotics were administered preoperatively to only four patients: three who were considered to have an infection on the basis of culture of fluid aspirated from the joint and one who was being managed for a urinary tract infection. Both joint fluid and tissue were obtained intraoperatively with swabs for aerobic and anaerobic cultures. Agar plates rather than broth were used for the cultures.
A positive intraoperative culture was the criterion for a diagnosis of infection. Comparison was made between the results of the analysis of the intraoperative frozen sections, the analysis of the permanent histological sections, and the intraoperative cultures. The numbers of positive and negative frozen and permanent sections were determined, and calculations were made to establish the rate of agreement between them. The results of other preoperative and intraoperative tests for infection (white-blood-cell count, erythrocyte sedimentation rate, C-reactive protein level, technetium and indium nuclear scans, culture of fluid aspirated from the joint, and intraoperative gram stains) were not analyzed in this study, as these tests were not uniformly performed.
Statistical analysis was performed to determine the sensitivity, specificity, and positive and negative predictive values of the intraoperative frozen sections. The two-tailed z test was used to determine significant differences among the various groups. The level of significance was set at p < 0.05.
Of the 175 patients, twenty-three had frozen sections that were considered positive (at least five polymorphonuclear leukocytes per high-power field). Of these twenty-three patients, eighteen had at least ten polymorphonuclear leukocytes per high-power field. Nineteen of the twenty-three patients who had at least five polymorphonuclear leukocytes per high-power field on analysis of the frozen sections had positive permanent sections (83 per cent agreement). All eighteen patients who had at least ten polymorphonuclear leukocytes per high-power field on analysis of the frozen sections had positive permanent sections (100 per cent agreement). Of the 152 patients who had less than five polymorphonuclear leukocytes per high-power field on analysis of the frozen sections, 144 also had negative permanent histological sections (95 per cent agreement).
Nineteen patients were considered to have an infection on the basis of positive intraoperative cultures. Staphylococcus aureus was the infecting organism in eight patients; coagulase-negative Staphylococcus, in six; Enterococcus and Pseudomonas, in two each; and Streptococcus viridans, in one patient. Twenty-seven patients had positive permanent histological sections, and eighteen of them had positive intraoperative cultures. The remaining 148 patients had negative permanent sections, and all but one had negative cultures. The permanent sections had a sensitivity of 94 per cent, a specificity of 98 per cent, a positive predictive value of 67 per cent, and a negative predictive value of 99 per cent.
Of the 152 patients who had less than five polymorphonuclear leukocytes per high-power field on analysis of the frozen sections, three had positive intraoperative cultures. This indicated that the frozen sections had been false-negative. None of the frozen sections were considered to be contaminated. All three patients had had a preoperative aspiration that had suggested the presence of infection; therefore, despite the negative frozen sections the treating physician staged the reimplantation, inserting the new prosthesis after a six-week course of antibiotics. Five patients had five to nine polymorphonuclear leukocytes per high-power field; all five had negative intraoperative cultures, and none had received antibiotics preoperatively. Sixteen of the eighteen patients who had at least ten polymorphonuclear leukocytes per high-power field had cultures that were positive for infection. Neither of the two patients who had negative cultures had received antibiotics preoperatively. Thus, seven patients had false-positive frozen sections on the basis of an index of five polymorphonuclear leukocytes per high-power field as compared with two patients who had false-positive results on the basis of an index of ten polymorphonuclear leukocytes per high-power field (Table I).
Seven patients had positive cultures that were considered by the treating physician to be due to a contaminant. These patients were considered not to have an infection when the statistical analysis was performed in this study. Corynebacterium, a common skin bacterium, grew on culture of specimens from four patients, and three patients had a weakly positive finding for coagulase-negative Staphylococcus on culture of one of eight specimens. Each of these patients had less than five polymorphonuclear leukocytes per high-power field on analysis of the frozen sections, and each was managed with a primary exchange arthroplasty. None of these patients had clinical signs of infection at a mean of twenty months (range, fourteen to twenty-eight months) after the operation.
The sensitivity of the frozen sections for detecting infection when it was present was 84 per cent regardless of whether an index of five or ten polymorphonuclear leukocytes per high-power field was used. The specificity of the intraoperative frozen sections increased from 96 per cent when an index of five polymorphonuclear leukocytes was used to 99 per cent when an index of ten polymorphonuclear leukocytes was used. There was a significant difference (p < 0.05) between the positive predictive value of the intraoperative frozen sections when an index of five polymorphonuclear leukocytes was used and the value when an index of ten polymorphonuclear leukocytes was used to predict infection (70 compared with 89 per cent). The frozen sections were equally effective in predicting the absence of infection (a negative predictive value of 98 per cent) regardless of whether an index of five or ten polymorphonuclear leukocytes per high-power field was used (Table II).
Primary exchange of the implant at the time of the index revision operation was performed in 153 patients. All but one of these patients had less than five polymorphonuclear leukocytes per high-power field on analysis of the frozen section. The other patient had five to nine polymorphonuclear leukocytes per high-power field, but a one-stage revision arthroplasty was performed because the complete preoperative workup (which included white-blood-cell count, erythrocyte sedimentation rate, gram stains, technetium and indium nuclear scans, C-reactive protein level, and aspiration of fluid from the joint as well as assessment of the gross appearance of the tissue at the arthrotomy) suggested failure of the prosthesis due to aseptic loosening. None of these patients had a positive culture. Twenty-two patients had an excisional arthroplasty; eighteen of them had delayed reimplantation of the prosthesis at a mean of 112 days after that procedure, and the remaining four had a resection arthroplasty or an arthrodesis of the knee as the definitive treatment.
Failure of total joint arthroplasty as the result of infection occurs less frequently than mechanical loosening; however, the implications for the result of treatment can be far greater with the former type of failure. The diagnosis of an indolent infection after a total hip or knee arthroplasty may be difficult to determine. The patient may report pain but clinically may have none of the cardinal signs of infection, such as fever, swelling, erythema, or drainage. Routine hematological studies, such as a white-blood-cell count and determination of the erythrocyte sedimentation rate, lack specificity; determination of the C-reactive protein level may be more valuable1,5,11,23,25. Imaging studies are often not conclusive. Indium-labeled leukocyte scans, while more specific for the detection of infection, do not have the same sensitivity as technetium scans16,17,21,22,24.
Cultures of joint aspirate before revision arthroplasty reveal an infection in 85 to 98 per cent of the instances in which one is present3. However, a relatively high rate of these cultures tend to be false-positive secondary to contamination with skin flora13,14,19,20,26. There has also been a high (10 to 20 per cent) prevalence of false-negative results, with most reports suggesting a sensitivity ranging from 60 to 75 per cent2,12-14,19. The alarming prevalence of false-positive and false-negative results on preoperative culture of aspirate as well as on intraoperative culture has provided an impetus for many physicians to reserve decisions regarding treatment until after histopathological review of specimens of tissue obtained at the time of the operation7,9,10,15.
In the absence of a universally accepted method for diagnosing infection in patients who have a loose total joint prosthesis, the analysis of intraoperative frozen sections has been proposed as a reliable and accurate method for guiding the surgeon at the time of operative exploration4,6,9,18,23. However, data quantifying the degree of acute inflammation that accurately indicates infection have been limited. The current investigation was designed to determine prospectively the value of intraoperative frozen sections for the identification of active infection during a revision hip or knee arthroplasty as well as to determine the number of polymorphonuclear leukocytes per high-power field that is the best indicator of active infection.
In 1973, Bullough as well as Charosky et al. described a qualitative histological difference in the soft-tissue response to acute or chronic infection as compared with particulate wear debris. Those authors documented an infiltration with polymorphonuclear leukocytes that was limited to acute infection. Later, in a series of thirty-six failed prostheses, Mirra et al. quantified the degree of inflammation in the presence of acute infection. They found more than five polymorphonuclear leukocytes per high-power field on analysis of tissue from around each of the fifteen implants that were associated with infection. Furthermore, they found no acute inflammatory response in association with the twenty-one prostheses that were not associated with clinical or bacteriological evidence of infection, including those associated with extensive particles of debris. Recently, Feldman et al. reported that use of the criterion of more than five polymorphonuclear leukocytes per high-power field had a sensitivity of 100 per cent and a specificity of 96 per cent for the detection of infection.
The conclusions of those studies were disputed by Fehring and McAlister, who questioned the reliability of analysis of frozen sections for guidance of intraoperative decision-making. They reported that analysis of frozen sections of specimens obtained in 107 revision total joint arthroplasties had a sensitivity of only 18 per cent, although this low value may have been due to sampling error caused by failure to collect tissue from representative areas. In their series, specimens for culture were not routinely obtained from the same samples of tissue that were sent for histological analysis. Additionally, Fehring and McAlister attributed greater importance to the over-all histological appearance than to an absolute number of polymorphonuclear leukocytes per high-power field.
The results of the current study suggest that the positive predictive value of analysis of intraoperative frozen sections is greatly enhanced if an index of ten rather than five polymorphonuclear leukocytes per high-power field is used. With use of the former index, although the sensitivity of only 84 per cent was not as high as that in earlier studies4,6,9, the specificity was 99 per cent and the positive predictive value was 89 per cent. There are three factors that could explain these results. First, the earlier studies were small and retrospective; therefore, bias in the selection of patients may have played a role. The present study was large and prospective, so the potential for selection bias was minimized. Second, in the current study, sampling error may have decreased the sensitivity, as samples of tissue may have been obtained indiscriminately at the time of the arthrotomy rather than granulation tissue or tissue that appeared infected having been obtained. Furthermore, false-positive or false-negative results may have occurred if the same tissue that was sent for histological analysis was not also cultured. Third, sampling error may have occurred in the laboratory. For preparation of frozen sections, two areas are chosen by the pathologist for staining. This is in contrast to the sequential histological analysis of multiple thin sections through the entire specimen that is routinely performed in the preparation of permanent sections of tissue. Therefore, the higher sensitivity of analysis of permanent histological sections is attributable at least in part to the fact that entire specimens are prepared and studied. Additionally, the paraffin preservative used in the preparation of permanent sections makes sectioning easier, leaves less artefact, and causes less destruction of the cell membranes than occurs in the preparation of frozen sections. As mentioned earlier, cells without well defined cytoplasmic borders were excluded from the count, and some of these cells may have been polymorphonuclear leukocytes. The sensitivity of analysis of the frozen sections may thus be enhanced if the surgeon is more selective in obtaining gross tissue, if multiple small samples are sent for analysis, and if more than two fields are analyzed histologically.
False-positive frozen sections (that is, those associated with negative intraoperative cultures) may be due to one of several causes: a fastidious organism that fails to grow in vitro, culture and frozen-section specimens that were taken from different areas, or the presence of a loculated infection8. Discrepancy between samples sent for histological analysis and those used for cultures, as well as bacteriological contamination of the specimen obtained for culture, may account for the false-negative frozen sections (that is, those associated with positive cultures).
In conclusion, the presence of at least ten polymorphonuclear leukocytes per high-power field on analysis of intraoperative frozen sections is highly suggestive of active infection during a revision hip or knee arthroplasty. Contrary to previous reports, five to nine polymorphonuclear leukocytes per high-power field, in appropriately analyzed specimens, is not necessarily consistent with infection. The yield will be greatly enhanced and false-negative results will be minimized with careful tissue-sampling at the time of both the operation and the histological examination. Analysis of intraoperative frozen sections, with use of the criteria described, is an accurate aid in the decision-making process that can be employed with confidence at the time of revision arthroplasty. Clearly, the pathologist must be well versed in the technique of preparation and the interpretation of frozen sections for this to be considered a truly reliable test for detection of acute infection. Additionally, the orthopaedic surgeon and the pathologist must have a close working relationship and a free exchange of information.
On the basis of these results, we perform resection arthroplasty and plan for staged reimplantation if there are at least ten polymorphonuclear leukocytes per high-power field. Primary exchange of the implant is performed if there are less than five polymorphonuclear leukocytes per high-power field. If there are five to nine polymorphonuclear leukocytes per high-power field, we tend to be guided by the over-all results of the comprehensive preoperative evaluation and the surgeon's impression at the time of the operation.
Aalto, K.; Osterman, K.; Peltola, H.; and |and |Rasanen, J.: Changes in erythrocyte sedimentation rate and C-reactive protein after total hip arthroplasty. Clin. Orthop.,184: 118-120, 1984.184118
1984
[PubMed]
Barrack, R. L., and |and |Harris, W. H.: The value of aspiration of the hip joint before revision total hip arthroplasty. J. Bone and Joint Surg.,75-A: 66-76, Jan. 1993.75-A66
1993
Brause, B. D.: Infections associated with prosthetic joints. Clin. Rheumat. Dis.,12: 523-536, 1986.12523
1986
Bullough, P. G.: Tissue reaction to wear debris generated from total hip replacements. In The Hip. Proceedings of the First Open Scientific Meeting of The Hip Society, pp. 80-91. St. Louis, C. V. Mosby, 1973.
Carlsson, A. S.: Erythrocyte sedimentation rate in infected and non-infected total-hip arthroplasties. Acta Orthop. Scandinavica,49: 287-290, 1978.49287
1978
[CrossRef]
Charosky, C. B.; Bullough, P. G.; and |and |Wilson, P. D., Jr.: Total hip replacement failures. A histological evaluation. J. Bone and Joint Surg.,55-A: 49-58, Jan. 1973.55-A49
1973
Cherney, D. L., and |and |Amstutz, H. C.: Total hip replacement in the previously septic hip. J. Bone and Joint Surg.,65-A: 1256-1265, Dec. 1983.65-A1256
1983
Fehring, T. K., and |and |McAlister, J. A., Jr.: Frozen histologic section as a guide to sepsis in revision joint arthroplasty. Clin. Orthop.,304: 229-237, 1994.304229
1994
[PubMed]
Feldman, D. S.; Lonner, J. H.; Desai, P.; and |and |Zuckerman, J. D.: The role of intraoperative frozen sections in revision total joint arthroplasty. J. Bone and Joint Surg.,77-A: 1807-1813, Dec. 1995.77-A1807
1995
Fitzgerald, R. H., Jr.; Nolan, D. R.; Ilstrup, D. M.; Van Scoy, R. E.; Washington, J. A., II; and |and |Coventry, M. B.: Deep wound sepsis following total hip arthroplasty. J. Bone and Joint Surg.,59-A: 847-855, Oct. 1977.59-A847
1977
Forster, I. W., and |and |Crawford, R.: Sedimentation rate in infected and uninfected total hip arthroplasty. Clin. Orthop.,168: 48-52, 1982.16848
1982
[PubMed]
Gristina, A. G., and |and |Kolkin, J.: Current concepts review. Total joint replacement and sepsis. J. Bone and Joint Surg.,65-A: 128-134, Jan. 1983.65-A128
1983
Insall, J. N.; Thompson, F. M.; and |and |Brause, B. D.: Two-stage reimplantation for the salvage of infected total knee arthroplasty. J. Bone and Joint Surg.,65-A: 1087-1098, Oct. 1983.65-A1087
1983
Levitsky, K. A.; Hozack, W. J.; Balderston, R. A.; Rothman, R. H.; Gluckman, S. J.; Maslack, M. M.; and |and |Booth, R. E., Jr.: Evaluation of the painful prosthetic joint. Relative value of bone scan, sedimentation rate, and joint aspiration. J. Arthroplasty,6: 237-244, 1991.6237
1991
[PubMed][CrossRef]
McDonald, D. J.; Fitzgerald, R. H., Jr.; and |and |Ilstrup, D. M.: Two-stage reconstruction of a total hip arthroplasty because of infection. J. Bone and Joint Surg.,71-A: 828-834, July 1989.71-A828
1989
Magnuson, J. E.; Brown, M. L.; Hauser, M. F.; Berquist, T. H.; Fitzgerald, R. H., Jr.; and |and |Klee, G. G.: In-111-labeled leukocyte scintigraphy in suspected orthopedic prosthesis infection: comparison with other imaging modalities. Radiology,168: 235-239, 1988.168235
1988
[PubMed]
Merkel, K. D.; Brown, M. L.; Dewanjee, M. K.; and |and |Fitzgerald, R. H., Jr.: Comparison of indium-labeled-leukocyte imaging with sequential technetium-gallium scanning in the diagnosis of low-grade musculoskeletal sepsis. A prospective study. J. Bone and Joint Surg.,67-A: 465-476, March 1985.67-A465
1985
Mirra, J. M.; Amstutz, H. C.; Matos, M.; and |and |Gold, R.: The pathology of the joint tissues and its clinical relevance in prosthesis failure. Clin. Orthop.,117: 221-240, 1976.117221
1976
[PubMed]
O'Neill, D. A., and |and |Harris, W. H.: Failed total hip replacement: assessment by plain radiographs, arthrograms, and aspiration of the hip joint. J. Bone and Joint Surg.,66-A: 540-546, April 1984.66-A540
1984
Patel, D.; Karchmer, A.; and Harris, W. H.: Role of preoperative aspiration of the hip prior to total hip replacement. In The Hip. Proceedings of the Fourth Open Scientific Meeting of The Hip Society, pp. 219-223. St. Louis, C. V. Mosby, 1976.
Pring, D. J.; Henderson, R. G.; Keshavarzian, A.; Rivett, A. G.; Krausz, T.; Coombs, R. R.; and |and |Lavender, J. P.: Indium-granulocyte scanning in the painful prosthetic joint. AJR: Am. J. Roentgenol.,147: 167-172, 1986.147167
1986
[PubMed]
Rand, J. A., and |and |Brown, M. L.: The value of indium 111 leukocyte scanning in the evaluation of painful or infected total knee arthroplasties. Clin. Orthop.,259: 179-182, 1990.259179
1990
[PubMed]
Rand, J. A., and |and |Fitzgerald, R. H., Jr.: Diagnosis and management of the infected total knee arthroplasty. Orthop. Clin. North America,20: 201-210, 1989.20201
1989
Reing, C. M.; Richin, P. F.; and |and |Kenmore, P. I.: Differential bone-scanning in the evaluation of a painful total joint replacement. J. Bone and Joint Surg.,61-A: 933-936, Sept. 1979.61-A933
1979
Shih, L. Y.; Wu, J. J.; and |and |Yang, D. J.: Erythrocyte sedimentation rate and C-reactive protein values in patients with total hip arthroplasty. Clin. Orthop.,225: 238-246, 1987.225238
1987
[PubMed]
Tietjen, R.; Stinchfield, F. E.; and |and |Michelsen, C. B.: The significance of intracapsular cultures in total hip operations. Surg., Gynec. and Obstet.,144: 699-702, 1977.144699
1977