Introduction
Aseptic loosening is the leading cause of revision after THA [23 4, 12, 14, 20 11 15 7, 20
The current consensus is that if two of five intraoperative cultures show growth of the same bacteria, that is evidence of PJI, whereas one positive culture or mixed growth of different bacteria generally is considered to be of questionable clinical relevance in the setting of clinically aseptic revision [1, 13, 15, 22 10
Therefore, we asked: (1) What is the clinical importance of bacterial growth in intraoperative tissue cultures taken during first-time revision of a clinically aseptic THA in terms of all-cause rerevision and rerevision for PJI specifically? (2) Is there a difference in outpatient antibiotic treatment patterns that is dependent on the number of positive intraoperative cultures?
Patients and Methods
Data Sources
The present register study was based on longitudinal data from four validated Danish national clinical quality and administrative registers. The study was approved by the Danish Health and Medicines Authority (case number 3-3013-1664/1), the Danish Data Protection Agency (journal number 16/15215), and the Danish Clinical Registries (RKKP).
The Danish Civil Registration System (Centrale Person Register [CPR]) was established in 1968 and contains information on all people living in Denmark, including immigrants. All 5.7 million Danish citizens are assigned a unique and unchangeable personal identification number known as the CPR number at birth or on immigration [3 21 17
The Danish Hip Arthroplasty Register (DHR) is a clinical quality database that gathers data from all THAs performed in Denmark, including both primary and revision THAs [9 23 16
Data were obtained on request for all THAs (identified by CPR number and side) that had been registered with a revision resulting from aseptic loosening. The data were acquired on May 17, 2017, after the monthly register update.
The Danish Microbiology Database (MiBa) was established in January 2010 and gathers microbiology reports from all departments of clinical microbiology in Denmark. All intraoperative tissue samples collected at Danish orthopaedic departments are analyzed at the nearest department of clinical microbiology in the same health region, and an electronic copy of the report is automatically sent to the MiBa. All requests and reports are kept using the CPR number as a patient identifier [25 2
The Danish National Prescription Registry contains individual-level data on prescriptions redeemed by Danish residents at public pharmacies since 2003. The register records the CPR number, prescribed drug identifier (Anatomical Therapeutic Chemical classification system [ATC]), date of prescription redemption, and pack size, among other variables. The indication for prescription is not recorded. The Danish National Prescription Registry has a high degree of completeness and validity; thus, it is extensively used in pharmacoepidemiologic research, including antibiotic use [18
Study Population
We examined the DHR for all reports of first-time revision surgery, which were recorded in the register as being the result of aseptic loosening of any prosthesis component from January 1, 2010, to May 15, 2016, excluding revisions performed as a result of other indications such as dislocation, fractures, and PJI (Fig. 1 ). Hip identification was obtained by combining the patient-unique CPR number with operated side and including both THAs if the patient had been revised bilaterally.
Fig. 1: Study flowchart describes cohort selection and the merging of databases.
THAs were excluded if there were errors in registering the CPR number, operated side, operation date, operation classification, or patient vital status (Fig. 1 ).
We identified 10,365 THAs from the DHR registered as revised resulting from aseptic loosening, of which 2305 were first-time aseptic revisions in 2243 patients performed within the study period (Fig. 1 ). The mean age was 73.4 ± 11.1 years; 56% of patients were women. The cohort was predominantly partial component revisions (67%) (Table 1 ).
Table 1.: Overall demography of the study cohort
DHR data on first-time clinically aseptic revisions and rerevision were linked using the CPR number, operation side, and operation date. DHR data were then linked with MiBa data containing national microbiologic results from growth of intraoperative cultures (Fig. 1 ). The study period was chosen to correspond with the introduction of the MiBa database on January 1, 2010, which maximized the sample size when the two databases were combined.
We defined unexpected positive intraoperative cultures as the presence of one or more positive cultures in a revision that the surgeon had defined as aseptic and reported to the DHR as such. First-time aseptic revisions were stratified according to the number of unexpected positive intraoperative cultures of the same bacterial genus, creating the following groups: culture-negative (reference population); one positive culture (this group included multiple positive cultures growing different bacteria [mixed growth]); and two or more positive cultures with the same bacteria. If no culture data were available, the revision was regarded as culture-negative. This classification was chosen because it reflects commonly used cutoffs in clinical decision-making.
Data on intraoperative cultures were obtained for 2090 of 2305 (91%) of first-time revisions. Of these first-time clinically aseptic revisions, 1855 (89%) had five samples collected, whereas 1938 revisions (93%) had between three and five samples collected as recommended by the Musculoskeletal Infection Society [15
Of the 2305 revisions reported as aseptic, 282 (12%) had bacterial growth in the intraoperative cultures. Of these, 170 (60%) had one positive culture (157) or had mixed bacterial growth (13). One hundred twelve revisions (40%) had two or more positive cultures growing the same bacteria (Table 2 ).
Table 2.: Microbiologic overview of first-time revisions (%)
The dominant bacteria in the presumed aseptic first-time revisions were recorded. This was defined as the most prevalent bacteria in the cultures, and in the case of different bacteria of equal prevalence, we recorded the bacteria most clinically relevant to PJI. Revisions with more than one bacterial genus grown from cultures were recorded as mixed-growth revisions.
Among all culture-positive revisions, coagulase-negative Staphylococcus was the dominant bacteria in 189 revisions (67%) followed by Propionibacterium species in 30 revisions (11%) and Staphylococcus aureus in 12 revisions (4%). This pattern was also present in revisions with one positive culture or mixed growth, coagulase-negative Staphylococcus being the dominant bacteria in 121 revisions (71%) (Table 2 ).
Followup
All first-time revisions were followed until rerevision, death, emigration, or end of the followup period, which was defined as 1 year. A 1-year followup period was chosen, because the majority of rerevisions resulting from PJI occur within the first year (approximately 80%-90%) [6 7
A rerevision was performed for 163 (7%) of the 2305 first-time clinically aseptic revisions within 1 year, whereas 87 patients (4%) died in the followup period without rerevision. None in the cohort emigrated during the 1-year followup period. In rerevisions, PJI was diagnosed in 43 patients (26%) (Fig. 1 ). Of the 163 THAs that were rerevised, 116 (71%) received a partial revision, whereas 47 (29%) were completely revised at their first-time revision. This representation of partially revised THAs among the rerevised THAs was not different from the overall cohort.
The DHR data were linked with the Danish National Prescription Registry at the individual level and reviewed for any antibiotic prescriptions (ATC category J01) redeemed within a 6-week period of the initial revision date. We recorded whether patients had been prescribed outpatient antibiotic treatment within this period, assuming that the patient had redeemed the prescribed medication. The prescribed drug was also recorded.
Within 6 weeks of surgery, 585 patients (25.4%) received an antibiotic prescription, of which 380 (16.5%) had possible relevance to PJI treatment.
Statistics
The primary outcome was risk of rerevision resulting from all causes within 1 year of first-time revision, including rerevisions resulting from PJI. The secondary outcome was risk of rerevision resulting from PJI alone. The relative risk for rerevision was estimated with the subdistribution hazard ratio with culture-negative first-time revisions as the reference [5
The relative risk for rerevision resulting from all causes was analyzed, treating death as a competing risk. The relative risk for rerevision resulting from PJI was analyzed, treating death and rerevisions resulting from other causes as competing risks. The proportional hazard assumption was assessed with the log–log plot and was not violated. We also calculated the cumulative incidence of outcome events in the groups; however, our outcomes were only analyzed using the subdistribution hazard ratio.
In our results, we report p values to describe whether the difference between groups is statistically significant. We also report the hazard rate ratio, hereafter referred to as the relative risk, to describe the magnitude of the difference. The relative risk is reported with 95% confidence interval; we remind readers that with this statistical approach, differences between groups can be statistically significant in some instances where the confidence intervals are overlapping.
The chi-square test was used to analyze the prevalence of patients with an antibiotic prescription including all antibiotics. This prevalence was subanalyzed, including only antibiotics relevant to PJI treatment in accordance with international literature and consensus [22, 26
We regarded a statistically significant difference as a p value < 0.05.
Results
Risk of Rerevision
The risk of all-cause rerevision was greater among first-time revisions with one positive culture, but not in the two or more positive culture group. In first-time revisions with one positive culture, the relative risk (RR) for all-cause rerevision was increased to 1.73 (95% confidence interval [CI], 1.07–2.80; p = 0.020) compared with the culture-negative revisions (Table 3 ). First-time revisions with two or more positive cultures did not have an increased risk of all-cause rerevision (RR, 1.52; 95% CI, 0.82–2.80; p = 0.180) compared with the culture-negative revisions.
Table 3.: Cumulative incidence and subdistribution hazard ratio for rerevision within 1 year resulting from all causes and PJI alone
First-time revisions in the one positive culture group also had a higher risk of rerevision for PJI specifically, but this was not the case in the two or more positive group. Revisions with one positive culture had a RR of 2.63 (95% CI, 1.16–5.96; p = 0.020) (Fig. 2 ). Additionally, when analyzing the revisions in the one positive culture group, excluding mixed-growth revisions, the RR for rerevision resulting from PJI remained increased to 2.44 (95% CI, 1.02–5.84; p = 0.045) compared with culture-negative revisions.
Fig. 2: Cumulative incidence of rerevision resulting from prosthetic joint infection within 1 year after first-time revision is reported as to the result of aseptic loosening: culture-negative group versus one positive culture group.
In revisions with two or more positive cultures growing the same bacteria, the RR for rerevision for PJI was 2.28 (95% CI, 0.81–6.43; p = 0.120).
All THAs in the one positive culture group that went on to be rerevised resulting from PJI had coagulase-negative Staphylococcus bacteria in the first-time revision.
Antibiotic Treatment
Outpatient antibiotic prescription was more frequent after revisions in the two or more positive culture group when compared with the culture-negative revisions. This was the case with regard to both all antibiotic prescriptions (51% [57 of 112 ] versus 24% [486 of 2023]; odds ratio [95% CI], 3.28 [2.23-4.81]; p < 0.001) and PJI-relevant prescriptions (50 of 112 [45%] versus 353 of 2023 [17%]; odds ratio [95% CI] 3.82 [2.58-5.63]; p < 0.001) (Table 4 ). In contrast, revisions with one positive culture did not have more patients with an antibiotic prescription when compared with the culture-negative revisions for either antibiotics overall (42 of 170 [25%] versus 486 of 2023 [24%]; odds ratio [95% CI] 1.04 [0.72-1.49]; p = 0.840) or PJI-relevant drugs specifically (36 of 170 [21%] versus 353 of 2023 [17%]; odds ratio [95% CI] 1.27 [0.86-1.87]; p = 0.220).
Table 4.: Prevalence of antibiotic prescription within 6 weeks after first-time revision
Discussion
Recognizing PJI in THA revisions is important to ensure appropriate patient care. Some THAs may be revised resulting from presumed aseptic loosening, but this diagnosis may be challenged by the presence of unexpected bacterial growth from intraoperative tissue samples. In revisions with ambiguous culture results such as one positive culture of several that were obtained during surgery, it can be difficult to choose the correct treatment for the patient; current consensus suggests that one positive culture is regarded as clinically irrelevant [15 Staphylococcus in the samples of the first-time revision, a common skin bacterium that could easily be mistaken for sample contamination.
Limitations
This study had several limitations. First, registry data are prone to miscoding and missing data; however, the DHR has been validated, finding completeness of both primary and revision surgery is very high (95%-98%) [23 7
Our primary outcome was assessed at 1 year to secure complete followup on the entire cohort while maximizing study power. Because this study evaluated the risk of rerevision from bacteria already present in the joint at the time of the first revision, we believe that 1 year is sufficient followup to capture the majority of revisions, because > 80% of PJIs have been shown to occur within the first year [6
Microbiologic data were available for 91% of first-time revisions and 82% of rerevisions. The lack of microbiologic data in the other revisions likely was caused by several factors: the surgeon deciding not to obtain intraoperative cultures, deeming the revision to be clearly aseptic; the surgeon forgetting to obtain samples; or because the cultures were lost during sampling or cultivation. Deciding to regard missing microbiology data as culture-negative revisions may have influenced our analysis. However, this decision resulted in some revisions in which bacteria could have been present as being classified in the reference group (culture-negative group). This would lead to overestimation of the risk of rerevisions in the reference population. Therefore, if the estimated RR for the THAs with positive cultures compared with the culture-negative THAs is skewed, it is most likely underestimated. All included THAs were deemed clinically aseptic by the reporting surgeon. Later unexpected growth of the same bacteria in two or more positive cultures would reclassify the revisions as not an aseptic loosening revision. However, these revisions were confined to their own subgroup and did not influence the analysis of the other groups. Some definitions of PJI regard growth of high-virulence bacteria such as S. aureus in only one culture as evidence of infection. In our cohort, such revisions were rare (five of 170) and none of these were later revised as a result of PJI.
Another concern was that most culture-positive revisions had only one positive culture or mixed growth (one positive culture group, 60%). It can be argued that PJI risk may be higher in mixed-growth revisions. In our study, such revisions numbered only 13 (8%) of the 170 THAs, and the risk of rerevision resulting from PJI remained increased when these were removed from the analysis. We chose to group the mixed-growth revisions together with the group of one positive culture, because they are both considered non-PJI according to most definitions, including the Musculoskeletal Infection Society consensus [13
We calculated RR of rerevision using competing risk analysis, which assumes that all included observations are independent. This is not entirely fulfilled when including bilateral THA; however, this has little practical consequence when studying revision [19
We reviewed the Danish National Prescription Registry data for 6 weeks after first-time revision surgery. Patients were considered as having received antibiotic treatment if just one prescription had been made, regardless of timing, pack size, or other variables. It is not possible to be certain if the prescriptions were revision-related, but we performed a subanalysis focusing on antibiotics of potential relevance to PJI treatment. However, prescription of such antibiotics may also be the result of superficial surgical site infection or other reasons, which may explain why antibiotic prescriptions were given to the culture-negative patients. We have no reason to believe that such circumstances are more frequent in the culture-positive revisions. The registry records prescriptions when they are redeemed at a pharmacy. It is possible that some antibiotic prescriptions were not redeemed by the patient, leading to underestimations of the prevalence of antibiotic treatment. We believe this to be rare and should not differ between the study groups.
Risk of Repeat Revision After Unexpected Positive Cultures
Unexpected growth in one intraoperative culture, including mixed growth with different bacteria, was associated with increased risk of all-cause rerevision and rerevision for PJI alone in this cohort. The dominant bacteria overall and in revisions with one positive culture was coagulase-negative Staphylococcus , which is a well-known organism in PJI [22 Staphylococcus was also the predominant bacteria in THAs that went on to be rerevised. These findings indicate that some cases of unexpected bacterial growth should likely be regarded as clinically important and not merely as contamination, although they would not be regarded as PJI according to current guidelines [15, 26 Fig. 2 ) support this, showing the highest increase in rerevision risk in the first 3 months, which is a common period for PJI revision [6 24
In our cohort, 12% of clinically aseptic revisions had positive cultures, which is in line with the study by Saleh et al. [20 15
A more recent study by Jacobs et al. [10
Antibiotic Treatment After Positive Cultures
We found that outpatient antibiotic prescription was more frequent after revision in patients with two or more positive cultures when compared with the culture-negative revisions, both with respect to all antibiotic prescriptions and to PJI-relevant prescriptions. This was in contrast to revisions with one positive culture who were treated with antibiotics as if they were culture-negative. This suggests that the surgeon did not believe that one positive culture merited a clinical intervention. This finding is interesting because it highlights the lack of importance attributed to unexpected positive cultures, both with regard to one positive culture which seems to be considered merely contamination, but also revisions with two or more positive cultures. Fifty-one percent of these patients had outpatient antibiotic treatment prescribed, which seems low considering that these can be defined as PJI and antibiotic treatment for 4 to 6 weeks has been recommended for such infections [22
The improved diagnosis of occult PJI in clinically aseptic THA is of great importance for future care of this large and growing patient group, where PJI is associated with increased morbidity, costs, and mortality [8 Staphylococcus . This emphasizes that unexpected bacterial growth with common bacteria may have clinical importance, even if only one of five biopsy cultures is positive [26
References
1. Atkins BL, Athanasou N, Deeks JJ, Crook DW, Simpson H, Peto TE, McLardy-Smith P, Berendt AR. Prospective evaluation of criteria for microbiological diagnosis of prosthetic-joint infection at revision arthroplasty. The OSIRIS Collaborative Study Group. J Clin Microbiol. 1998;36:2932-2939.
2. Bank S, Søby KM, Kristensen LH, Voldstedlund M, Prag J. A validation of the Danish microbiology database (MiBa) and incidence rate of
Actinotignum schaalii (Actinobaculum schaalii) bacteraemia in Denmark. Clin Microbiol Infect. 2015;21:1097.e1-4.
4. Fernandez-Sampedro M, Salas-Venero C, Farinas-Alvarez C, Sumillera M, Perez-Carro L, Fakkas-Fernandez M, Gomez-Roman J, Martinez-Martinez L, Farinas MC. Postoperative diagnosis and outcome in patients with revision arthroplasty for aseptic loosening. BMC Infect Dis. 2015;15:232.
5. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496-509.
6. Gundtoft PH, Overgaard S, Schonheyder HC, Moller JK, Kjaersgaard-Andersen P, Pedersen AB. The 'true' incidence of surgically treated deep prosthetic joint infection after 32,896 primary total hip arthroplasties: a prospective cohort study. Acta Orthop. 2015;86:326-334.
7. Gundtoft PH, Pedersen AB, Schonheyder HC, Overgaard S. Validation of the diagnosis 'prosthetic joint infection' in the Danish Hip Arthroplasty Register. Bone Joint J. 2016;98:320-325.
8. Gundtoft PH, Pedersen AB, Varnum C, Overgaard S. Increased mortality after prosthetic joint infection in primary THA. Clin Orthop Relat Res. 2017;475:2623-2631.
9. Gundtoft PH, Varnum C, Pedersen AB, Overgaard S. The Danish Hip Arthroplasty Register. Clin Epidemiol. 2016;8:509-514.
10. Jacobs AME, Benard M, Meis JF, van Hellemondt G, Goosen JHM. The unsuspected prosthetic joint infection : incidence and consequences of positive intra-operative cultures in presumed aseptic knee and hip revisions. Bone Joint J. 2017;99:1482-1489.
11. Kamme C, Lindberg L. Aerobic and anaerobic bacteria in deep infections after total hip arthroplasty: differential diagnosis between infectious and non-infectious loosening. Clin Orthop Relat Res. 1981;154:201-207.
12. Nelson CL, McLaren AC, McLaren SG, Johnson JW, Smeltzer MS. Is aseptic loosening truly aseptic? Clin Orthop Relat Res. 2005;437:25-30.
13. Parvizi J, Jacovides C, Zmistowski B, Jung KA. Definition of periprosthetic joint infection: is there a consensus? Clin Orthop Relat Res. 2011;469:3022-3030.
14. Parvizi J, Suh DH, Jafari SM, Mullan A, Purtill JJ. Aseptic loosening of total hip arthroplasty: infection always should be ruled out. Clin Orthop Relat Res. 2011;469:1401-1405.
15. Parvizi J, Zmistowski B, Berbari EF, Bauer TW, Springer BD, Della Valle CJ, Garvin KL, Mont MA, Wongworawat MD, Zalavras CG. New definition for periprosthetic joint infection: from the Workgroup of the Musculoskeletal Infection Society. Clin Orthop Relat Res. 2011;469:2992-2994.
16. Pedersen A, Johnsen S, Overgaard S, Soballe K, Sorensen HT, Lucht U. Registration in the Danish Hip Arthroplasty Registry: completeness of total hip arthroplasties and positive predictive value of registered diagnosis and postoperative complications. Acta Orthop Scand. 2004;75:434-441.
17. Pedersen CB. The Danish Civil Registration System. Scand J Public Health. 2011;39:22-25.
18. Pottegard A, Schmidt SAJ, Wallach-Kildemoes H, Sorensen HT, Hallas J, Schmidt M. Data Resource Profile: The Danish National Prescription Registry. Int J Epidemiol. 2017;46:798-798f.
19. Ranstam J, Karrholm J, Pulkkinen P, Makela K, Espehaug B, Pedersen AB, Mehnert F, Furnes O. Statistical analysis of arthroplasty data. II. Guidelines. Acta Orthop. 2011;82:258-267.
20. Saleh A, Guirguis A, Klika AK, Johnson L, Higuera CA, Barsoum WK. Unexpected positive intraoperative cultures in aseptic revision arthroplasty. J Arthroplasty. 2014;29:2181-2186.
21. Schmidt M, Pedersen L, Sorensen HT. The Danish Civil Registration System as a tool in epidemiology. Eur J Epidemiol. 2014;29:541-549.
22. Tande AJ, Patel R. Prosthetic joint infection. Clin Microbiol Rev. 2014;27:302-345.
23. The Danish Hip Arthroplasty Register. Annual Year Rapport 2016. 2016. Available at:
http://danskhoftealloplastikregister.dk/wp-content/uploads/2015/11/DHR-%C3%A5rsrapport-2016.pdf . Accessed November 19, 2018.
24. Trampuz A, Piper KE, Jacobson MJ, Hanssen AD, Unni KK, Osmon DR, Mandrekar JN, Cockerill FR, Steckelberg JM, Greenleaf JF, Patel R. Sonication of removed hip and knee prostheses for diagnosis of infection. N Engl J Med. 2007;357:654-663.
25. Voldstedlund M, Haarh M, Mølbak K; MiBa Board of Representatives. The Danish Microbiology Database (MiBa) 2010 to 2013. Euro Surveill. 2014;19. pii: 20667.
26. Zimmerli W, Trampuz A, Ochsner PE. Prosthetic-joint infections. N Engl J Med. 2004;351:1645-1654.
