Subclinical Abnormal Coagulation in Periprosthetic Joint Infection: A Retrospective Cohort Study

Background(cid:0)Periprosthetic joint infection (PJI) is a serious complication of total joint arthroplasty and often indicate disastrous outcomes. However, the change of coagulation prole in PJI patients has not been explored up to now. Therefore, we performed a single-center retrospective cohort study to determine: 1) the coagulation prole in PJI patients 2) the diagnostic ecacy of coagulation prole for PJI diagnosis based on the MSIS criteria. Methods: Between 2016 January and 2018 December, a total of 371 patients receiving joint revisions were included in this cohort study. The corresponding medical records were scrutinized to establish the nal diagnosis of PJI according to the 2014 MSIS criteria. The difference of coagulation prole between PJI and aseptic loosening patients was analyzed. Moreover, receiver operating characteristic curves were used to determine the proper sensitivity and specicity of coagulation makers. Results: The levels of APTT, D-dimer, plasma brinogen, INR and Platelet Count in PJI group were signicantly higher than that in non-PJI group(P<0.05). The AUCs of plasma APTT, plasma D-dimer, plasma brinogen and platelet count for PJI diagnosis were0.625(95%CI:(0.543,0.706)),0.731(95%CI: (0.656,0.806)),0.831(95%CI:(0.771,0.890)) and 0.733(95%CI:(0.660,0.805)), respectively. Moreover, the coagulation prole was combined by logistic model and the corresponding AUC was0.865(95%CI: (0.812,0.918)). Conclusions: Despite relatively normal coagulation prole, PJI patients suffer from subclinical abnormal coagulation compared to non-PJI patients. The coagulation prole (APTT, INR, plasma brinogen, platelet count, D-dimer) in PJI patients is different from that in non-PJI patients signicantly. And the coagulation prole can play a role in PJI diagnosis.


Introduction:
Total joint arthroplasty (TJA) is a successful surgery during the last century because it relieves pain and improves the quality of life for patients with advanced joint diseases [1]. Periprosthetic joint infection (PJI), indicating unfavorable outcomes, is one of the most disastrous complications after total joint arthroplasty (TJA) [2]. And coagulation system is a complex system and abnormal coagulation in perioperative TJA patients is often associated with serious complications such as deep venous thrombosis (DVT), surgical site infection (SSI), deep infection and increased blood loss [3]. Moreover, coagulopathic patients often have delays in surgery and cause aggravated anxiety to both surgeons and patients [3].
Application of coagulation pro le in the diagnosis of PJI is emerging recently. Many studies reveal that the levels of plasma brinogen and D-dimer in PJI patients are higher than that in non-PJI patients and these markers can play a role in PJI diagnosis [4][5][6]. Besides, D-dimer was introduced into the 2018 ICM criteria by Parvizi [7][8][9]. Plasma brinogen and D-dimer are indictors of coagulation and brinolytic system, respectively. The abnormal brinogen and D-dimer in PJI patients may indicate the presence of abnormal coagulation. In [8], Arjun Saxena e.t suggested that PJI can cause abnormal coagulation and more than one half of patients with PJI may have abnormal coagulation even without being exposed to anticoagulation agents. But the abnormal coagulation in his study was de ned when INR > 1.12. However, this de nition can lead to bias because INR only re ect extrinsic pathway of coagulation cascade and corresponding intrinsic pathway wasn't evaluated in this study [8]. Besides, several studies also revealed that the use of plasma brinogen and D-dimer for the diagnosis of PJI [10,9,11,12]. Recently, the use of platelet count for PJI diagnosis was proposed by Pei F [11]. Despite that these studies suggested the change of coagulation pro le in PJI patients, these studies only re ected limited aspects of coagulation system. A comprehensive coagulation pro le in PJI patients haven't been explored and built up to now.
In order to comprehensively explore the change of coagulation pro le in PJI patients and establish a PJI diagnostic tool based on the preoperative coagulation tests, we conducted a single-center retrospective cohort study to determine 1) the change of coagulation pro le in PJI patients 2) the diagnostic e ciency of coagulation pro le for PJI based on the MSIS criteria.

Materials And Methods:
Inclusion and exclusion criteria: Between 2016 January and 2018 December, a total of 371 suspected PJI patients who underwent joint revisions were included in this study as initial patients. The exclusion criteria were as follows: 1) internal xation implantation 2) patients who were exposed to anticoagulation agents (NSAID, corticosteroid hormone, clopidogrel e.g.) or trauma within 2 weeks before revisions 3) coronary stents or lter implantation 4) periprosthetic fractures 5) prosthetic dislocation. Finally, 332 patients (PJI patients and non-PJI patients) were included in this study. The detailed information on this process was summarized in Figure 1.

De nition:
Suspected PJI was considered when a patient suffered from acute or persistent rest pain, swelling, redness, warmth around the joints, elevated erythrocyte sedimentation rate (ESR) and/or C-reactive protein (CRP) level, or implant failure within 5 years after total joint arthroplasty without any reasonable explanation.
The diagnostic criteria and data collection: The medical records of these patients were scrutinized and the nal diagnosis of PJI was based on the 2014 MSIS criteria (table 4) [13]. Besides, the chart review were performed by the rst authors carefully and following data about coagulation pro le were extracted: activated partial thromboplastin time (APTT), prothrombin time (PT), international normalized ratio (INR)thrombin time (TT), antithrombin 3 (AT3), prothrombin activity (PTA), plasma D-dimer, plasma brinogen and plasma Ca+. Moreover, following information was extracted: the age, sex, body mass index (BMI), ASA scores, the comorbidities of liver diseases, kidney diseases, heart diseases, lung disease, diabetes, and in ammatory joint diseases.

Statistical analysis:
The continuous variables were described as means and standard deviations and a t test was used to assess the comparison if normal distribution of continuous variables was detected. Otherwise, rand sum test was used and corresponding medians were calculated. Dichotomous data described as frequencies and percentages were compared by the chi-squared test. P<0.05 indicated statistical signi cance. Receiver operating characteristic curves (ROC) were used to determine the relationship between the sensitivity and speci city of coagulation pro le. Logistic regression was used to build the diagnostic model based on signi cantly up-regulated coagulation indicators (APTT, plasma brinogen, D-dimer, and platelet count). Moreover, Youden's index was used to determine optimal cut-off values. SPSS (IBM; version 26.0) was used to perform all statistical analysis. Results: 1. Demographic characteristics: The median age in the PJI and non-PJI groups was 63 and 61years, respectively, and no signi cant differences found between the two groups. Similarly, no signi cant differences were found in the percentages of females, liver diseases, kidney diseases, lung diseases, heart diseases, smoking and drinking. However, the level of BMI and the percentage of hip circumference were signi cantly different between the PJI and non-PJI groups. The details were shown in Table 1.  The diagnostic e ciency of coagulation pro le was evaluated and depicted in ROC curves (Fig.2). The AUCs of APTT, D-dimer, plasma brinogen and platelet count were 0.625(95%CI: (0.543,0.706)),0.731(95%CI:(0.656,0.806)),0.831(95%CI:(0.771,0.890)),0.733(95%CI:(0.660,0.805)), respectively. The optimal cut-off value of APTT for PJI diagnosis was 37.25 and the corresponding sensitivity and speci city were 0.637 and 0.637 respectively. The optimal predictive cutoff value of Ddimer was 0.745 and the corresponding sensitivity and speci city were 0.853 and 0.537 respectively. The optimal cut-off value of platelet count for PJI diagnosis was 245, and the corresponding sensitivity and speci city were 0.676 and 0.712 respectively. The optimal cut-off value of plasma brinogen for PJI diagnosis was 4.13 and the corresponding speci city and sensitivity were 0.755 and 0.825 respectively.
Moreover, the coagulation pro le was combined by logistic model (appendix 1) and the corresponding AUC was 0.865(95%CI:(0.812,0.918)). The details about the diagnostic e ciency of coagulation pro le and serum markers for PJI diagnosis were summarized in Table 3. Table 3 The In this study, we found that the coagulation pro le in PJI patients was different from that in non-PJI patients and PJI patients can suffer from subclinical abnormal coagulation pro le. In consistent with previous reports, some coagulation markers such as plasma D-dimer and brinogen show promising PJI diagnostic values. Besides, we also revealed a few new markers (APTT, INR and platelet count) for PJI diagnosis from the coagulation pro le in PJI patients. Then we build a logistic model based on these 3 markers and the ROC analysis suggested this model was promising in PJI diagnosis.
PJI indicating that infection occurs to the prosthesis and surrounding tissues is one of the most disastrous complications after TJA and its pathogens are known to impair their host by secreting endotoxin and exotoxin which can stimulate CD8 + T cells and antigen-presenting cells (APC) to produce various cytokines [8]. These cytokines (IL-1, TNFα, and IL-6) can disrupt normal coagulation cascade and subsequently cause abnormal coagulation pro le [14][15][16]. However, a review of literature suggests that the coagulation pro le in PJI patients remain unknown. This study evaluated the change of coagulation pro le in PJI patients.
In the presence of chronic infection, large amounts of endothelial cells can be disrupted and following cytokines and tissue factors will be released into the blood. Then coagulation system and following brinogen system can be activated [17]. Eventually, the body will delete excessive coagulation factors such as antithrombotic factors, TFPI (tissue factor pathway inhibitor) for hemostasis [18,19]. Then, following coagulopathy may appear in PJI patients.
The levels of APTT, INR, D-dimer, plasma brinogen, platelet count in the PJI group is signi cantly higher than that in non-PJI group. Despite relatively higher levels of APTT and INR, the levels of these two markers is still within clinical reference values. It suggests that PJI patients are suffering from subclinical abnormal coagulation. The levels of APTT and INR in PJI group is signi cantly higher than that in non-PJI group but the level of TT between the two groups is comparative. This condition reveals that the change of coagulation cascade in PJI group is mainly ascribed to the change of intrinsic pathway and extrinsic pathway instead of the common pathway (Fig. 3). Therefore, the coagulation factors participating in the two pathways of coagulation may be different between the two groups. However, this hypothesis needs approving further. And subsequent studies should focus on the change of coagulation factors in PJI patients.
The subclinical coagulation change may have potential in uence on increased blood loss in PJI patients [20]. Elevated APTT and INR indicate impaired coagulation system and can cause increased joint effusion after revision. This is consistent with our clinical observation that patients after PJI revisions often have larger drainage volume and longer drainage tube retention time than non-PJI patients. However, the larger joint effusion volume can increase the likelihood of PJI recurrence conversely. It can be paradoxical when doctors try to hold the balance between the abnormal coagulation and VTE in PJI patients after revision. Besides, vitamin K and FFP which is often used in the treatment of patients with coagulopathies are associated with severe reactions such as allergy.
[21] Therefore, correcting the abnormal coagulation in PJI patients may be a challenging work. And the association between abnormal coagulation pro le and the outcomes of PJI need to be explored further.
The level of APTT and INR in PJI patients is signi cantly higher than that in non-PJI patients. It means that PJI patients suffer from hypocoagulation compared to non-PJI patients. In [8], Parvizi reveled similar results. Despite a fact that most coagulation indictors were still within clinical reference values, we do hold the opinion that the subclinical coagulation pro le still causes some abnormalities. The use of regional anesthesia in these patients can increase the risk of epidural or spinal hematoma formation theoretically. And hypocoagulation may impair incision healing after surgery.
In this study, we also evaluated the use of coagulation pro le for PJI diagnosis. We found that APTT, platelet count, plasma brinogen and D-dimer can play a role in the diagnosis of PJI. The use of D-dimer for the diagnosis of PJI have been introduced into the 2018ICM criteria and the application of plasma brinogen and platelet count in PJI diagnosis has also been revealed by several studies [10,11]. But the accuracy of APTT has not been studied in suspected PJI patients. In our research, we found that plasma brinogen performed better than D-dimer and plasma brinogen when these markers were evaluated. It is consistent with previous studies and suggests that plasma brinogen can be a better marker than D-dimer [10,12]. Besides, although platelet count and the level of APTT had inferior diagnostic value compared with CRP, ESR, plasma brinogen and D-dimer, they can be useful for the diagnosis of PJI.
To further assess the performance of coagulation pro le in PJI diagnosis, a logistic model was built based on signi cantly up-regulated coagulation indicators in PJI patients (APTT, plasma brinogen, Ddimer, and platelet count). And this model performed better than CRP and ESR in the diagnosis of PJI. This result suggested that the coagulation pro le can be used as a screening tool in suspected PJI patients because this is a routine test before surgery and the corresponding cost is relatively low. Besides, this test can be performed in most hospitals. It makes this test become a simple screening tool for PJI diagnosis in primary hospital.
This study also has several limitations. First, even though the relationship between the coagulation pro le and PJI was revealed by our study, we can't assess the change of certain coagulation factors because of limited medical information. Second, this was a retrospective study performed in a single center so the study design and absent in medical records can lead to selection bias. Besides, it is hard for us to discern the causal link between PJI and abnormal coagulation pro les. Third, some patients with systemic diseases (kidney disease, liver diseases and systemic in ammatory diseases e.g.) were not excluded from this study [22]. These conditions can in uence the coagulation pro le of PJI patients. However, the heterogeneous cohorts can also provide a more realistic clinical situation for the evaluation of the coagulation pro le in PJI patients.

Conclusions:
PJI patients suffer from subclinical abnormal coagulation. The coagulation pro le in PJI patients is different from that in non-PJI patients and can serve as a method for PJI diagnosis. However, the relationship between the subclinical hypocoagulation and clinical outcomes of PJI patients need to be explored further.  Table   Table 4 The 2014 MSIS criteria for PJI diagnosis: (1) Elevated serum C-reactive protein (CRP) AND erythrocyte sedimentation rate (ESR) (2) A single positive culture  The ROC Curve of Different Coagulation Markers for PJI Diagnosis