What Is the Impact of Automated Synovial Cell Counting on Different Aseptic Causes and Periprosthetic Conditions Associated With Revision THA?

Abstract

Background

Studies have suggested that automated synovial cell counting may overestimate the white blood cell (WBC) count, resulting in false positive tests when evaluating patients for the possibility of periprosthetic joint infection (PJI) after THA. However, associations between WBC counts high enough to mimic PJI in patients whose arthroplasties are not infected but rather are experiencing a variety of aseptic problems—including but not limited to metallosis, polyethylene wear, and recurrent dislocation—have not, to our knowledge, been adequately addressed. In addition, there is a lack of analyses about the polymorphonuclear percentage (PMN%) when assessed by automated analyzers in this context.

Questions/purposes

In the context of different indications for aseptic revision and different periprosthetic hip pathologic findings, we asked: (1) What were the synovial WBC count levels, and what proportion of values were above the 2018 International Consensus Meeting (ICM) cutoff (3000 cells/μL)? (2) What were the synovial PMN% levels, and what proportion of values were above the 2018 ICM (70%)?

Methods

We retrospectively studied the preoperative cell count analyses of synovial fluid in patients who underwent revision THA for aseptic reasons at our tertiary referral arthroplasty center between January 2015 and December 2017. We considered all revisions performed on patients during that time potentially eligible, and after prespecified exclusions were applied (exclusions mainly included 15% [197 of 1306] sporadic missing data and 12% [155 of 1306] insufficient synovial fluid obtained in the aspirate), a total of 702 patients undergoing revision THA for aseptic reasons remained for the final analysis. As far as we know, no patients underwent re-revision for PJI at a mean follow-up of 46 ± 11 months, which tends to confirm our impression that indeed these hips did not have PJI. Cell count analyses were conducted using an automated analyzer. Clinical findings, preoperative radiographs, and surgical reports––confirmed by available histologic results––were used to establish diagnoses. We evaluated these hips considering the recommendations of the 2018 ICM (WBC count of 3000 cells/μL and PMN% of 70%) to see what proportion of them would have been characterized as likely having PJI on basis of those cutoff values. The mean WBC count for the entire cohort was 2120 ± 2395 cells/μL. The mean PMN% for the entire cohort was 36% ± 22%.

Results

Compared with aseptic loosening and recurrent dislocation, polyethylene wear had the highest mean WBC count (3817 ± 3711 cells/μL; p < 0.001). Of the investigated periprosthetic conditions, wear-induced synovitis had the highest value (4464 ± 3620 cells/μL; p < 0.001). Considering the ICM threshold, polyethylene wear showed the highest proportion of WBC counts above 3000 cells/μL among the indications for aseptic revision (60% [25 of 42]; p < 0.001). Of the periprosthetic conditions, wear-induced synovitis showed the highest proportion beyond the ICM cutoff (60% [50 of 83]; p < 0.001). The mean PMN% for aseptic causes ranged between 28% and 44% without differences among them (p = 0.12). Patients with metallosis had the highest mean PMN% of the periprosthetic conditions investigated (45% ± 25%; p = 0.007). Regarding the ICM threshold, metallosis resulted in the largest proportion of patients with a PMN% above 70% (21% [10 of 47]; p = 0.003), and that for wear-induced synovitis was 6% (5 of 83; p = 0.42) and for osteolysis was 3% (1 of 33; p = 0.51). There were no differences among aseptic loosening, recurrent dislocation, and polyethylene wear in terms of the proportion above 70% among the aseptic revision causes.

Conclusion

Using automated cell counting, we found that WBC counts differ widely across indications for aseptic revision THA, and a high proportion of patients who underwent revision THA for aseptic reasons had WBC counts above the commonly used threshold of 3000 cells/μL. However, the PMN% was much less affected in several common indications for aseptic revision THA, making this measure more reliable for interpreting aspiration results using an automated analyzer. Based on the data distributions we observed, manual counting techniques might be considered in particular among patients with polyethylene wear, metal-on-metal bearing surfaces, or suspected metallosis. However, an elevated WBC count alone, observed using an automated analyzer in the context of polyethylene wear, should not be considered to be strongly suggestive of PJI, since that finding occurred so commonly among patients without infection.

Level of Evidence

Level III, diagnostic study.

Introduction

In revision arthroplasty, joint aspiration remains the gold standard for diagnosing or ruling out periprosthetic joint infection (PJI) preoperatively, and increased accuracy can be achieved by using synovial fluid investigations during PJI treatment rather than serum markers and tissue-based tests [2, 3]. Although the white blood cell (WBC) count is among the most used synovial diagnostic markers globally, it demonstrates less specificity than other synovial markers [11]. This has been attributed to the association between certain indications for aseptic revision THA and elevated synovial WBC count [3, 11]. Previous studies have explored an altered synovial WBC count and/or differential count in patients with adverse local tissue reactions, aseptic loosening, and dislocation after THA [1, 4, 5]. However, those studies investigated aseptic revision THA, and they did not make it clear whether the cell counting method was automated, manual, or a combination of the two. The generally high proportion of false-positive WBC counts has resulted in concerns about the use of automated synovial cell counting, particularly for THA [7]; however, associations with different indications for aseptic revision THA have not, to our knowledge, been addressed. Furthermore, there is a lack of analyses considering the impact of automated analyzers on the polymorphonuclear percentage (PMN%). Thus, elevated synovial WBC counts and PMN% can be a source of confusion during the diagnostic workup before revision THA in which a surgeon must decide whether PJI is present or absent.

Therefore, in the context of different indications for aseptic revision, and different periprosthetic hip pathologic findings, we asked: (1) What were the synovial WBC count levels, and what proportion of values were above the 2018 International Consensus Meeting (ICM) cutoff (3000 cells/μL)? (2) What were the synovial PMN% levels, and what proportion of values were above the 2018 ICM (70%)?

Patients and Methods

Study Design and Setting

We retrospectively studied the preoperative cell count analyses of synovial fluid in patients who underwent revision THA for aseptic reasons at our tertiary referral arthroplasty center between January 2015 and December 2017. Our institution’s practice is to perform a preoperative joint aspiration on all patients undergoing scheduled (nonemergent) aseptic revision arthroplasty. Revisions for known PJI were excluded based on the ICM criteria [14, 17].

Participants

We identified 1306 revision THAs in which hip aspirations were performed. An aspirate was excluded if it did not have a cell count analysis available or featured a dry tap or insufficient synovial fluid. We included 73% (954 of 1306) of the synovial samples. We then excluded revision procedures without any prior surgery on the hip under investigation, patients who had undergone an open biopsy within the previous 6 months, patients with previous surgery for PJI pertaining to the index hip, patients with any positive culture results from intraoperatively obtained samples, and procedures that were followed by any surgery for infection on the index hip (Fig. 1). The most frequent exclusions were for sporadic missing data (15% [197 of 1306]) and insufficient synovial fluid obtained in the aspirate (12% [155 of 1306]). In total, 702 aseptic patients remained for analysis. As far as we know, no patients underwent re-revision for PJI at a mean follow-up duration of 46 ± 11 months.

Fig. 1.

This flowchart shows the patients who were included in this study.

Patients’ Baseline Data

Sixty-two percent (434 of 702) of patients were women (Table 1). Clinical findings, preoperative radiographs, and surgical reports––confirmed by available histologic results––were used to establish diagnoses. Causes of revision THA included aseptic loosening (68% [475 of 702]), repeated THA dislocation (16% [113 of 702]); polyethylene wear (6% [42 of 702]), periprosthetic fracture (2% [11 of 702]), implant fracture (2% [13 of 702]), impingement of the iliopsoas tendon (2% [12 of 702]), elevated metal ion level (1% [6 of 702]), and miscellaneous causes (4% [30 of 702]), such as cup-taper impingement, abductor deficiency, leg length discrepancy, heterotopic ossification, liner dislocation, and ceramic fracture. Regardless of the main revision cause, we categorized each patient according to whether they had one of three main, concurrent periprosthetic conditions: wear-induced synovitis (defined as the gross granulomatous inflammatory appearance of periprosthetic tissue and the joint capsule with the partial formation of necrotic tissue), metallosis (defined as the gross observation of metallic, stained periprosthetic tissues), and osteolysis (defined as any obvious acetabular or femoral osteolytic bone loss).

Table 1.

Characteristics of the entire study cohort (n = 702)

Characteristic	Value
Age in years	69 ± 11
Female	62 (434)
BMI in kg/m2	27 ± 5
Preoperative CRP in mg/L	4 ± 8
Time in yearsa	11 ± 9
Synovial WBC count in cells/µL	2120 ± 2395
Synovial PMN%	35 ± 22
Synovial mononuclear %	64 ± 22

Data are presented as mean ± SD or % (n).

^aBetween prior arthroplasty and the index revision; CRP = C-reactive protein.

Data Collection and Variables

For each revision procedure, the following data were retrieved using available electronic medical records: preoperative and intraoperative diagnoses, results of preoperative aspirations, preoperative C-reactive protein levels, microbiologic and histopathologic results of periprosthetic tissue samples taken intraoperatively, prosthetic data (bearing surface and cementless or cemented implants), patient comorbidities (rheumatoid arthritis, diabetes mellitus, and gout), and demographics (age, sex, and BMI).

Aspiration Protocol and Cell Count Analysis

At the hospital where the study was performed, all patients scheduled for any revision THA also underwent aspiration. For the current cohort, antibiotic therapy was stopped for at least 14 days before aspiration. The hips were aspirated under sterile conditions using fluoroscopy to ensure appropriate intra-articular positioning of the aspiration needle. Aspirates were sent for microbiologic culturing, cell count analysis (WBC count and differential analysis, including PMN% and mononuclear %), and α-defensin immunoassay. Given that the α-defensin immunoassay was not an established, standard synovial investigation method during the first year of the study period, these data were not available for approximately one-third of the aspirate samples. Saline lavage was not used in any patients in the cohort. Cell count analyses were conducted using an automated analyzer (Sysmex XN-9000, Sysmex Europe). Based on fluorescence flow cytometry, analyses were performed using the body fluid mode. Previous studies have demonstrated the excellent performance of this analyzer for evaluating cell count and differentiation in body fluids, including synovial fluid [13, 16].

The mean WBC count for the entire cohort was 2120 ± 2395 cells/μL. The mean PMN% for the entire cohort was 36% ± 22%.

Primary and Secondary Study Outcomes

We evaluated all hips in light of the recommendations of the 2018 ICM (WBC count of 3000 cells/μL and PMN% of 70%) [15] to see what proportion of them would have been characterized as likely having PJI on basis of those cutoff values. Our primary study outcome was WBC count assessment, and our secondary outcome was the PMN%. We also investigated whether there were any identifiable factors associated with elevated WBC counts or PMN%.

Ethical Approval

Ethical approval for this study was obtained from the Ethik-Komission der Arztekammer Hamburg, Hamburg, Germany (WF-115120).

Statistical Analysis

The number, mean, median, and SD were used to report metric values in the descriptive analysis. To verify the normal distribution of the data, we applied the Shapiro-Wilk test. We compared groups using a t-test. Either the chi-square or Fisher exact test was applied to compare the distribution frequency of categorical variables. Otherwise, the Mann-Whitney U or Kruskal-Wallis test was used. All tests were two-tailed. A bivariate analysis was performed to identify factors associated with elevated WBC counts and PMN%.

Results

WBC Count and Percentage Above 3000 cells/μL

In terms of aseptic causes, the diagnosis with the highest mean WBC count was polyethylene wear (polyethylene wear 3817 ± 3711 cells/μL; p < 0.001) compared with aseptic loosening (2005 ± 2292 cells/μL) and recurrent dislocation (1744 ± 1703 cells/μL) (Table 2). Wear-induced synovitis (4464 ± 3620 cells/μL; p < 0.001) and osteolysis (3565 ± 3555 cells/μL; p = 0.002) had higher mean WBC counts compared with metallosis (3103 ± 4063 cells/μL; p = 0.37) among the periprosthetic conditions we investigated. For the 2018 ICM threshold, we found that 17% (116 of 702) of the aspirate samples demonstrated a WBC count greater than 3000 cells/μL, and 3% (18 of 702) had both an elevated WBC count and a PMN% count greater than 70%. Polyethylene wear showed the highest proportion of WBC counts above 3000 cells/μL (60% [25 of 42]; p < 0.001) compared with aseptic loosening (16% [75 of 475]) and recurrent dislocation (14% [16 of 113]) (Fig. 2) for aseptic causes of revision. Among the periprosthetic conditions, the proportion of WBC counts higher than the ICM cutoff for wear-induced synovitis was 60% (50 of 83; p < 0.001), the proportion for osteolysis was 45% (15 of 33; p < 0.001), and the proportion for metallosis was 32% (15 of 47; p = 0.021) (Fig. 2). Younger patients (mean age 67 ± 11) were more likely to have WBC count > 3000 cells/μL compared with older patients (mean age 70 ± 11; p = 0.003) among the variables we analyzed. Likewise, metal-on-metal bearings were associated with higher proportion of WBC count > 3000 cells/μL (56% [9 of 16]) compared with ceramic-on-polyethylene (19% [69 of 359]) and metal-on-polyethylene (18% [56 of 318]; p = 0.005) (Table 3).

Table 2.

WBC count of different aseptic revision causes of THA and periprosthetic conditions

Variable	Cells/µL
Revision cause (n)
Aseptic loosening (475)	2005 ± 2292
Recurrent dislocation (113)	1744 ± 1703
Polyethylene wear (42)	3817 ± 3711
Implant fracture (13)	1965 ± 1347
Iliopsoas impingement (12)	3520 ± 2796
Periprosthetic fracture (11)	1200 ± 796
Elevated metal ions (6)	1821 ± 1167
Periprosthetic condition (n)
Wear-induced synovitis (83)	4464 ± 3620
Osteolysis (33)	3565 ± 3555
Metallosis (47)	3103 ± 4063

Data presented as mean ± SD.

Fig. 2.

A-B Box plots demonstrating the synovial white blood cell (WBC) count (cells/μL). Boxes represent the interquartile range, and the horizontal lines inside the boxes represent the median. Outliers and extreme values are presented as spots. The broken line refers to the International Consensus Meeting (ICM) threshold. (A) Indication for aseptic revision: Note that only the median value of polyethylene wear (3466 cells/μL) is above the ICM cutoff. (B) Periprosthetic condition: Note that only the median value of wear-induced synovitis (3921 cells/μL) is above the ICM cutoff.

Table 3.

Results of the bivariate analysis of investigated factors for any association with elevated WBC count (> 3000 cells/μL) and PMN% (> 70%)

Variable	WBC < 3000 cells/μL	WBC > 3000 cells/μL	p value	PMN% < 70	PMN% > 70	p value
Age in years	70 ± 11	67 ± 11	0.003	69 ± 11	68 ± 11	0.34
Sex  Female (n = 434)  Male (n = 268)	82 (355) 80 (213)	18 (79) 20 (55)	0.42	91 (397) 92 (247)	9 (37) 8 (21)	0.78
BMI	27 ± 5	27 ± 4	0.38	27 ± 5	27 ± 4	0.98
Diabetes mellitus  Yes (n = 73)  No (n = 629)	82 (60) 81 (508)	18 (13) 19 (121)	0.86	90 (66) 92 (578)	10 (7) 8 (51)	0.62
Rheumatoid arthritis  Yes (n = 36)  No (n = 666)	83 (30) 81 (538)	17 (6) 19 (128)	0.45	94 (34) 92 (610)	6 (2) 8 (56)	0.19
Gout  Yes (n = 12)  No (n = 690)	83 (10) 81 (558)	17 (2) 19 (132)	0.91	83 (10) 92 (634)	17 (2) 8 (56)	0.26
Fixation Cemented/partially cemented  (n = 429)  Cementless (n = 273)	83 (355) 78 (213)	17 (74) 22 (60)	0.1	91 (392) 92 (252)	9 (37) 8 (21)	0.74
Bearing surfacea  CoP (n = 359)  MoP (n = 318)  MoM (n = 165)  CoC (n = 3)	81 (290) 82 (262) 44 (7) 100 (3)	19 (69) 18 (56) 56 (9) 0 (0)	0.005	92 (329) 93 (294) 75 (12) 100 (3)	8 (30) 7 (24) 25 (4) 0 (0)	0.11

Data are presented as mean ± SD or % (n).

^aTotal number = 696, as bearing surfaces could not be specified in six hips; CoP = ceramic-on-polyethylene; MoP = metal-on-polyethylene; MoM = metal-on-metal; CoC = ceramic-on-ceramic.

PMN% and Percentage Above the 70% Threshold

The mean PMN% for aseptic causes ranged between 28% ± 25% for periprosthetic fracture and 44% ± 25% for iliopsoas impingement (Table 4). Metallosis was associated with the highest PMN% of the periprosthetic conditions investigated (mean 45% ± 25%; p = 0.007) compared with wear-induced synovitis (mean 34% ± 23%; p = 0.41) and osteolysis (mean 35% ± 22%; p = 0.81). Of all aspirates, only 6% (40 of 702) had a higher PMN% than the 2018 ICM threshold. Patients with metallosis also had the largest proportion of PMN% above 70% (21% [10 of 47]; p = 0.003); for wear-induced synovitis, it was 6% (5 of 83; p = 0.42), and for osteolysis, it was 3% (1 of 33; p = 0.51) (Fig. 3). Among aseptic reasons, the proportion of patients with a PMN% higher than the ICM threshold ranged between 7% (31 of 475) for aseptic loosening and 25% (3 of 12) for iliopsoas impingement (Fig. 3). However, there were no differences (p = 0.12) between the large groups, including aseptic loosening, recurrent dislocation (10% [11 of 113]), and polyethylene wear (14% [6 of 42]).

Table 4.

Polymorphonuclear percentage (PMN%) of different aseptic revision causes of THA and periprosthetic conditions

Variable	PMN%
Revision cause (n)
Aseptic loosening (475)	35 ± 21
Recurrent dislocation (113)	34 ± 23
Polyethylene wear (42)	42 ± 23
Implant fracture (13)	31 ± 25
Iliopsoas impingement (12)	44 ± 25
Periprosthetic fracture (11)	28 ± 25
Elevated metal ions (6)	41 ± 24
Periprosthetic condition (n)
Wear-induced synovitis (83)	34 ± 23
Osteolysis (33)	35 ± 22
Metallosis (47)	45 ± 25

Data presented as mean ± SD.

Fig. 3.

A-B Box plots demonstrating the percentage of polymorphonucleocytes (PMN%) of the synovial white blood cell (WBC) count for (A) aseptic revision and (B) periprosthetic condition. Boxes represent the interquartile range, and the horizontal lines inside the boxes represent the median. The broken line refers to the International Consensus Meeting threshold.

Discussion

Despite evidence that altered synovial fluid WBC counts vary among different aseptic indications for revision THA [1, 4-7, 13], synovial WBC counts and PMN% values continue to produce confusion before revision THA, resulting in unnecessary testing, repeated aspirations, and even misdiagnoses. Therefore, we sought to identify which aseptic causes and periprosthetic conditions are commonly associated with elevated WBC counts or PMN% using automated cell counting in revision THAs for definitively aseptic reasons. We found that a relatively high proportion of our sample (17% [116 of 702]) had a WBC count above the ICM threshold (> 3000 cells/μL). Among the aseptic causes investigated, we observed that polyethylene wear had a higher proportion of WBC counts above 3000 cells/μL than aseptic loosening and recurrent dislocation. Meanwhile, given that wear-induced synovitis and periprosthetic osteolysis commonly occur secondary to polyethylene wear, these factors demonstrated a higher proportion of WBC counts above 3000 cells/μL than metallosis. However, metallosis was associated with a higher proportion of PMN% values above the ICM threshold than the other periprosthetic hip conditions. Overall, the proportion of PMN% values above the ICM-recommended threshold was much lower than the proportion of WBC counts above the ICM threshold across the total cohort. Based on this, we believe surgeons should only consider WBC and differential counts to be suggestive of infection, for purposes of indicating further diagnostic testing, if both the WBC count is elevated above 3000 cells/μL and the differential is greater than 70% PMN using automated counting, and not consider it suggestive of infection if only one or the other of those metrics is elevated.

Limitations

Our findings should be interpreted considering some important limitations. First, the study had a small sample of some groups, such as the iliopsoas impingement group, which is a relatively rare cause of revision arthroplasty. Therefore, the results in these groups should be interpreted cautiously, as a larger or more varied sample might have yielded different findings; however, we do not believe this to be a disqualifying shortcoming because the investigated large groups are the major aseptic indications for revision THA. Second, the absence of cell count analyses or dry taps meant that 27% (352 of 1306) of patients were initially excluded. This was because of the absence of a firmly established standard aspiration protocol during the initial study period. Further exclusion criteria were applied to ensure reliable synovial cell count results. Nonetheless, we believe that the patients ultimately included represent a considerable proportion of the total sample. Third, given the possibility of concurrent findings and indications for aseptic revision THA, there is always some uncertainty in making these diagnoses. To reduce this risk, three experienced revision arthroplasty surgeons (HA, AA, MR) were involved in defining the final diagnosis and the main periprosthetic diagnoses based on all available data.

WBC Count and Percentage Above 3000 cells/μL

According to our findings, wear-induced synovitis was associated with higher WBC counts than metallosis. To the best of our knowledge, polyethylene wear-induced increases in synovial WBC counts have not been well documented in connection with revision THA. Polyethylene wear particles activate the cellular response and cause a third-body reaction that is mainly mediated by macrophages [18]. The severity of consequent synovitis has been correlated with the extent of polyethylene damage in THA and TKA [10, 12], and wear debris-related synovitis was associated with increased WBC counts in an analysis of 54 revised TKAs [15]. The automated analyzer we used might have yielded falsely elevated values because of agitated periprosthetic particles or crystals, with one large cohort study indicating a 34.3% false-positive rate (> 3000 cells/μL) for THA compared with 10.1% for TKA [7]. The correlation with the corresponding manual count was consequently moderate for THA, strong for TKA, and perfect for native knees [7]. Therefore, interpretations of the WBC counts in THA should be undertaken with caution, and a manual analysis can serve as a valuable tool in cases of doubt, especially in patients with metal bearing surfaces or corrosion [19]. However, this requires substantial experience, high costs, and more time, and might be associated with a lack of interlaboratory coordination [13].

Elevated WBC counts in patients with iliopsoas impingement could be the result of the mechanical irritation caused by a protruded acetabular component or massive peritendineal scarring that promotes an enhanced inflammatory reaction. Recurrent THA dislocation is a further indication for aseptic revision THA that contributes to changes in the synovial cell count [1]. In a small study investigating 29 THA dislocations, there were increased synovial WBC counts without PMN% elevation when hips underwent aspiration within 1 day after closed reduction [1]. Meanwhile, our patients with dislocation after THA had results comparable to those of patients with aseptic loosening. Our patients underwent revision because of recurrent dislocation, and aspirations were performed within a few weeks before revision, regardless of the timing of the last dislocation. Aseptic loosening was the most common cause of revision THA in our cohort, and these patients had a mean WBC count of 2025 cells/μL. Sixteen percent (75 of 475) of patients had values that were higher than the ICM cutoff, which might be associated with periprosthetic diagnoses. In a previous study, a WBC count below 2500 cells/μL combined with a percentage of lymphocytes above 10% could be used to diagnose aseptic loosening with considerable accuracy [5]. A study investigating the synovial cell count before the revision of 433 TKAs, including infected TKAs, revealed different values for different indications for aseptic revision TKA, producing different cutoff levels for diagnosing PJI according to the cause of aseptic revision. This study found the optimal cutoff was 4697 cells/μL for periprosthetic fracture and 2104 cells/μL for component malposition [6]. The average WBC counts observed in that study were generally lower than those of the groups in our cohort; however, the synovial cell count analysis method was not reported [6]. Therefore, an elevated WBC count alone, observed using an automated analyzer, should not be considered strongly suggestive of PJI, since that finding occurred commonly among patients without infection but who had polyethylene wear. Also, we believe that future studies should determine the best methods and techniques of cell counting when reporting on synovial cell analysis in the setting of THA.

PMN% and Percentage Above the 70% Threshold

With regard to the PMN%, polyethylene wear was associated with the highest average values, joining aseptic loosening and recurrent dislocation as the principal causes for revision. However, patients with iliopsoas impingement had the largest elevations beyond the ICM threshold (3 of 12 patients). As previously discussed, debris-induced synovitis and the peritendineal inflammatory reaction might be responsible in these patients. In terms of current ICM cutoffs, our results suggest that the PMN% is much less affected by the automated synovial analysis approach than the total WBC count (6% versus 17% for the total WBC count). Hence, the combined evaluation of WBC count and PMN% should always be a part of the approach to PJI when using automated analyzers (before choosing repeat aspiration). Nonetheless, 3% of our study cohort had higher WBC counts and PMN% values and would receive 5 points according to the 2018 ICM criteria, indicating a high likelihood of PJI (patients receiving 3 to 5 points were considered inconclusive) [17]. Notably, negative results were observed for all other minor criteria, including serum C-reactive protein levels less than 10 mg/L, the absence of purulence, and negative histologic and intraoperative culture results. At a minimum follow-up duration of more than 2 years, none of these patients that we know of underwent re-revision for PJI.

Although one multicenter study considered the optimal cutoffs for diagnosing hip PJI as a WBC count of 3966 cells/μL and a PMN% value of 80%, the researchers observed a gray area [9]. Using a cutoff of 3000 cells/μL for the WBC count and 80% for the PMN%, false-positive results were found for 33% of THAs with corrosion or metal-on-metal bearing surfaces. However, manual counting yielded an accuracy of 98% for the WBC count and 90% for the PMN% [19]. For THA caused by metallosis, 31.9% of patients had WBC counts higher than 3000 cells/μL, although a lower proportion of patients (21.3%) had PMN% values higher than the 2018 ICM cutoff (70%). Metallosis promotes metal-induced synovitis, and metal-on-metal bearing surfaces can be complicated in patients with adverse local tissue reactions. Adverse local tissue reactions are associated with increased values for synovial WBC count and PMN%, challenging the approach to treating PJI [4]. Thus, the 2018 ICM acknowledged that the recommended cutoff levels might not be accurate in patients with certain periprosthetic conditions, such as adverse local tissue reactions [17]. Importantly, adverse local tissue reactions include a combination of periprosthetic tissue necrosis, pseudotumor, lymphocyte-dominated vasculitis-associated lesions, osteolysis, and metallosis [8].

Conclusion

Using automated cell counting, we found that WBC counts differ widely across indications for aseptic revision THA, and a high proportion of patients who underwent revision THA for aseptic reasons had WBC counts above the commonly used threshold of 3000 cells/μL. However, the proportion of patients with elevated PMN% was much lower than that with an elevated WBC count above the ICM cutoffs, making this measure more reliable for interpreting aspiration results using an automated analyzer. Although polyethylene wear was the most common indication for aseptic revision THA associated with a high WBC count, periprosthetic metallosis was associated with higher PMN% values, often above the ICM cutoffs. Therefore, manual counting techniques might be considered in patients with polyethylene wear, metal-on-metal bearing surfaces, or suspected metallosis. However, an elevated WBC count alone, observed using an automated analyzer in those with polyethylene wear, should not be considered as strongly suggestive of PJI, since that finding occurred commonly among patients without infection. Further research might consider the value of the WBC or differential count as a preoperative diagnostic tool using automated cell counting during the early stages of polyethylene wear after THA.