PJI with Sepsis, a Complex Emergency Situation with Prodromal Stage

Introduction: With a considerable morbidity, mortality and a high financial impact, PJI has been considered as one of the most serious complications after arthroplasty. Special awareness must be given to these patients due to a life threatening systemic PJI related complication, sepsis. A prompt detection and adequate medical management of this situation is crucial for avoiding unfavorable outcomes. Methods: Between January 2011 and December 2012, we retrospectively examined adult patients who met PJI-SIRS criteria. Medical history search for patients’ anamnesis, surgical times, laboratory-microbio-logical findings and success rates was performed. Results: Twenty patients were enrolled in this study with a mean age of 71.35 years. Men population was more commonly affected with 55% of the cases. The knee was mainly the affected joint with 52%. A severe systemic disease was present in 80% of the cases. A sickness sensation and pain were the most common prodromal symptoms found 4.9 days before the admission to the hospital. In 85.7% of the cases the etiological agent was identified, with Staphylococcus aureus as the most common. Two-stage surgery treatment was performed with a time interval between admission and the first operation of 5.45 h. The mean time during the first and second operation was of 23 minutes and 117 minutes respectively. A reduction on microbiological


Introduction
With higher life expectancy and new therapeutic approaches for joint diseases, the general population has increasingly higher expectations of mobility in old age. This leads to a steady increase in the number of joint replacement operations, which is expected to rise by 137% and 601% for total hip and knee endoprostheses by 2030 [1]. Although endoprostheses are successful surgical procedures, with more than 95% of survivors undergoing surgery within a follow-up period of 10 years [1], the number of periprosthetic joint infections (PJI) will also increase. PJI is a devastating and serious complication after arthroplasty. PJI is considered a rare event that occurs in about 2% of primary and 4% of revision endoprostheses [2] [3] and its incidence is lower than that of aseptic loosening [4]. Nevertheless, it can lead to more serious complications for endoprostheses, such as prolonged hospitalization, multiple operations, functional impairment and even death, for which a five-year mortality rate of 21.12% for total hip endoprostheses and 14.4% for total knee endoprostheses after 3.8 years has been reported [5] [6].
There are different classification schemes for PJI. The classification into acute and chronic PJI is the most frequently used and published [7]. The most important criterion of these classifications is the condition and formation of a bacterial biofilm. It is assumed that acute PJI begins < 4 weeks after prosthesis implantation or, in the case of hematogenic etiology, has a maximum symptom duration of less than 3 weeks. In these cases, the biofilm is not yet fully developed. These acute PJIs are most frequently triggered by relatively highly virulent microorganisms such as Staphylococcus aureus and Gram-negative bacteria. In contrast, chronic PJI (low grade) are delayed postoperative infections, i.e. >than 4 weeks after prosthesis implantation or, in the case of hematogenic genesis, a symptom duration of more than 3 weeks [7] [8]. The biofilm age of this type of infection is usually mature. The main germs of chronic PJI are relatively low-virulent microorganisms, such as coagulase-negative staphylococci or propionibacterium acnes [7] [8].
The most life-threatening complication for patients with PJI is the sepsis-disease.
Patients with PJI may develop bacteremia, which can trigger a systemic inflammatory response (SIRS) through certain bacterial components such as endotoxins recognized by immune cells [9]. The combination of infection and SIRS is called sepsis. In the sepsis spectrum, SIRS is the first complication step, which also includes severe sepsis and septic shock with organ dysfunction [10]. It has been described that mortality in patients with severe sepsis and septic shock is about 50%, which is almost 6 times higher than in patients without sepsis [11] [12].
Early detection and rapid adequate treatment of the early stage of sepsis in patients with PJI is crucial for the prognosis and prevention of higher mortality.

Study Group
We retrospectively collected data from 20 adult patients diagnosed with PJI who met the SIRS criteria [13], who were admitted and treated at our clinic between January 2011 and December 2012. This represented 7.5% of all 286 patients with PJI at our clinic during this period. SIRS was defined according to the criteria of the German Sepsis Society. In short: fever or hypothermia, tachycardia, tachypnea, leukocytosis or leukopenia were used as criteria parameters. In the presence of a positive blood culture and 2 of these parameters, the patients were defined as patients with SIRS. In the presence of a negative blood culture, patients must have 4 of these parameters to define a SIRS. Exclusion criteria consisted in patients under 18 years old and an incomplete medical record. The retrospectively evaluated data of our study was anonymized and the information was only accessible to the treatment team.

Therapy Concept
The treatment concept for the patients included emergency diagnostics and a two-stage operation concept. The diagnosis included a complete anamnesis with recording the type and onset of the prodromal symptoms such as joint redness, heat, swelling, pain, fever and feeling of illness. The two-stage surgical concept included an initial immediate emergency operation with arthrotomy of the joint, synovectomy and debridement. This surgery also included joint lavage, the use of antibiotic labels and the placement of a drainage tube for subsequent daily sterile joint lavage, until a second delayed and planned surgical revision. At this time, the microbiological culture results of the primary diagnosis and primary surgery had to be available. The second surgery included debridement, joint lavage and the use of absorbable antibiotic sponges. Only in case of prosthesis loosening or the microbiological findings of a multi-resistant pathogen we saw the indication for prosthesis explantation.

Samples and Laboratory Tests
In the diagnostic phase a preoperative joint puncture was performed for synovial fluid samples. A synovial fluid WBC count and a microbiological culture were performed. This phase also included radiographs to identify prosthesis loosening and the type of prosthesis. Peripheral blood samples were taken to determine WBC count, inflammation markers such as CRP, coagulation tests, kidney and liver function parameters, electrolytes and procalcitonin levels. In addition, blood cultures for aerobic and anaerobic microorganisms were prepared at patient admission. C-reactive protein and WBC count in peripheral blood were measured daily throughout the patient's hospital stay. Microbiological cultures were also established from synovial fluid/tissue samples collected during the first and second surgery.

ASA Classification
In addition, the classification of the physical status of the American Society of Anesthesiologists (ASA) [14], which was performed by a specialist on the day of the patient's anesthesia treatment, was added to our database. In short, ASA 1 is considered a healthy person, ASA 2 is a patient with a mild systemic disease, ASA 3 is a patient with a severe systemic disease, ASA 4 is a severe systemic disease that poses a constant threat to the patient's life, ASA 5 is a moribund person who is not expected to survive without surgery, and an ASA 6 classification, i.e. a person declared brain dead whose organs have been removed for donor purposes.

Statistical Analyses
All the data was analyzed using descriptive statistics, all patients narratives and clinical information during the anamneses, diagnosis and treatment was converted into structured data). The Mann Whitney non-parametric Gaussian distribution test was used between 2 groups of data. Fisher exact test was used to analyze microbiological culture data using Prism GraphPad 8 (San Diego, California).

Results
Demographic and clinical characteristics of the patients As shown in Table 1 a total of twenty patients were enrolled in this study with a follow up time mean of 9.4 months (range from 3 -27 months). The mean age of the patients was of 71.35 years old with a range from 45 -89 years. Men were more commonly affected with a 55% of the cases (11 patients) compared to women with a 45% (9 patients). A slight difference was found between ages of men and women which had a mean of 69.6 years (49 -82) and 73.6 years (45 -89) respectively. The most common affected joint was the knee with a 52% of the cases (10 patients,), followed by hip with a 43% and shoulder 5%. The vast majority of the patients presented a mono articular PJI (95%) and only 1 patient    A two-step operation treatment focused on a rapid first operation The time interval between the admission and the first operation was of 5.45 h (2.5 -12 h). On the other hand, the time gap between the first and second operation was of 8 days (4 -12 days). The mean duration of the first and second operation was of 23 minutes (13 -52) and 115.7 minutes (71 -208) respectively.
In those patients were conservation of the prosthesis was opted a mean duration of operation of 83.3 minutes was observed, in contrast to those where the prosthesis needed to be removed which lasted a mean duration of 148 minutes. The time gap between the removal and replacement of the prosthesis had a mean of 6.3 weeks (Figure 2).  Probable prodromal signs or symptoms were presented by 90% of the patients before admission, leaving 10% without any prodromal variable to measure (Table 2). A sickness sensation and pain were the most common prodromal symptoms patients presented with a 90% and 85% respectively. These was followed by joint swelling in a 70% of the cases. Fever and joint warmth were present in 60% while the less common prodrome was the joint redness with a 25%. The time interval of onset of illness before inpatient treatment and the appearance of these prodromal variables was of 4.9 days (1 -10 days).

Microbiological cultures % of positiveness decreases 3 times after the first rapid operation
In the 85.7% of the cases (17 patients The most common isolated bacteria in synovial fluid samples belonged to Staphylococcus spp. and Streptococcus spp. with a 52.2% and 28.6% respectively, leaving another 4.8% for other bacteria species. Staphylococcus aureus was responsible for more than the half of the cases of the Staphylococcus spp followed by Staphylococcus epidermidis. On the other hand, Streptococcus pneumoniae was the most common Streptococcus spp isolated in this group (Table 3).  Table 4.
CRP serum levels decreased after first rapid operation Serum C-reactive protein showed statistically significant lower levels the next day after the first rapid surgery (P = 0.0137) as shown in     respectively were observed during the first day of hospitalization. The next day after the first surgery the range was from 116 -289 mg/L with a median value of 198 mg/L and a mean of 195 mg/L. After the second surgery these CRP serum levels decreased almost 2 times (P = 0.0042) with a range from 18.10 to 194 mg/L and a median and mean values of 98.52 mg/L and 111 mg/L respectively and almost 3 times when compared to the admission day levels (P = 0.0005). Peripheral blood WBC count didn't show a statistically significant difference through the days of the patient's hospitalization time (P > 0.05). Figure 5 shows the decrease tendency of CRP concentrations through patients medical management.
Normal body temperature after second day of hospitalization During the admission, patients presented a body temperature mean of 38.60˚C which decreased after the first rapid surgery to 37.32˚C (P = 0.0181) and 36.76˚C (P = 0.0006) on the second and third day of hospitalization respectively. From day 3 on patients body temperature remained without any statistically significant change ( Figure 6).

Half of the patients had an infection resolution without prostheses removal
Our results showed that 100% of the patients belonging to the group treated without a protheses removal (10 patients) had a successful management of the infection with a follow up mean time of 10.8 months (range 5 -27). On the other hand, patients whom underwent to a 2-stage prosthesis exchange (8 patients) had a also successful treatment rate of 100% in controlling the infection with a follow up mean time of 7.8 months (range 3 -18). One patient (5%) died due to multiorgan disfunction syndrome related to multiple PJIs and a Girdlestone resection arthroplasty was needed also in 5% of the cases (Table 5).

Discussion
Although PJI occurs in less than 5% of all endoprostheses [2] [3], it remains a severe, devastating and serious complication. PJI with sepsis in particular can also have systemic effects with increased mortality [15]. For this reason, the early and correct detection of early signs, symptoms and laboratory parameters is of crucial importance, not only to detect the joint infection early, but also to prevent higher mortality due to sepsis.
In our retrospective study we were able to show that on average 4.9 days before sepsis the first symptoms already appeared. Leading among the symptoms were joint pain and an unspecific feeling of illness. Our results showed that joint pain occurred in 85% of patients, which is consistent with other studies that showed its occurrence in almost 80% -100% of cases [16] [17] [18].
However, different from these studies, pain was not the most common manifestation the patients were presenting in our research. Most of our patients (90%) described a general sickness sensation as the most common clinical manifestation before their arrival to the hospital. We were able to demonstrate that sepsis is preceded by a prodromal stage that should be used therapeutically and diagnostically.
Consistent with other studies the majority of our patients (55%) were men, which population has been considered as a non-modifiable risk factor for developing PJI [19] [20] [21]. Although no scientific explanation has yet been proposed for this phenomenon a reasonable explanation could be job or comorbidities related which makes this population more susceptible for the infection. Our results showed a slight not significant difference in the age of the patient's affection, which was older in women (73.6 years) compared to men (69.3 years). One must take into consideration that although the importance of age in many other diseases, age alone has not been considered as risk factor for developing PJI [22].
Of the 3 different joints affected, the knee was most frequently affected with 52% of cases, followed by the hip with 42% and the shoulder with 5%. This is consistent with previous studies that showed PJI rates between 0.5% -2% for all knee joint prostheses, followed by almost 1% for all hip and shoulder prostheses in less than 1% [23] [24].
Probably due to the fact that the knee is the largest joint in the human body, it is also the most commonly affected joint in PJI, which is also due to the greater mobility, larger prosthetic surface area and less soft tissue coverage. However, there could also be a bacterial cause. Certain bacteria, such as Staphylococcus aureus, have surface components that bind or attach to extracellular matrices, such as fibrinogen, collagen and elastin [25] [26]. These components, which are considered determinants of virulence [27], may also facilitate the attachment of the bacteria to implanted materials that are coated with host proteins after insertion [28]. This attachment to a foreign implanted material could develop a microbial community embedded in a protective extracellular matrix called biofilm  [29].
As our results show, a positive blood culture could only be obtained in 25% of cases. Although bacteremia is a known consequence of PJI, these cultures are not often obtained or considered as part of the diagnostic criteria of PJI [30] [31]. It should be noted that if positive, they may reduce the success rate of PJI treatment [32] and even lead to death, as one of our patients who presented a positive MRSA blood culture and unfortunately died in septic shock. However, the role of blood cultures in PJI diagnosis is an issue that requires further research.
It is well known that Gram-positive bacteria are the most common bacteria involved in PJI. Staphylococcus aureus represents the most common etiological agent with more than the half of the cases [8]. This is accordingly to our results were this etiological agent represented the 52.2%. However other study showed Staphylococcus epidermidis as the most common Gramm-positive cause of the infection [33] which in our case was the second most common. Our data showed that Streptococcus spp were present in a 28.6%, which has been described to represent approximately 10% of the cases [8].
Something to consider was the percentage of positive results in our synovial fluid/tissue cultures from preoperative joint puncture, first intraoperative and second intraoperative surgical specimens. In 85.7% of cases, identification of the etiologic pathogen from preoperative puncture was possible, in 14.3% of cases not. A negative culture is a relatively common finding in the diagnosis of PJI and is present in 0% -40% of cases [34] [35]. It is important to note that the presence of this microbiological result does not always mean an incorrect PJI diagnosis. This can have various reasons, such as prior use of antibiotics, which was the case in our patients who received systemic antibiotic treatment before sampling. However, many other reasons may cause a negative result, such as inadequate culture conditions for not frequent bacterial joint infections, the low presence of the etiologic pathogen, which may already have been eliminated but whose antigens, toxins, or superantigens could cause an inflammatory response, and clinical manifestations in the patient [36] [37]. In this case, new diagnostic alternatives are being investigated, such as multiplex PCR, which has been described as a reliable early diagnostic tool in patients with PJI and SIRS, where early pathogen identification is cardinal [38].
Another common possibility, though particularly in chronic PJI, of negative cultures is the presence of biofilms, where the need for a different type of culture may require procedures for accurate diagnosis, such as sonication [39] [40].
Compared to preoperative puncture, the number of positive cultures from the samples from the first surgery decreased by 10.7% from 85.7% to 75% and the number of negative cultures increased accordingly from 14.3% to 25%. A possible explanation could be that after diagnosis and preoperative sampling, antibiotics were given to patients, which may have influenced other cultures even after a few hours of administration. On the other hand, the results of the cultures at the time of sampling were completely reversed at the second surgery compared to the first surgery, resulting in a 75% and 25% number of negative and positive cultures, respectively.
Several risk factors, including an increased ASA score, have been associated with an increased risk of infection at the surgical site. This is particularly true for patients with an ASA score ≥ 3 [23] [41], as our results show, where the majority of our patients (80%) had an ASA classification score of 3 (severe systemic disease) on arrival at our hospital.
Even today, PJI treatment still poses a great challenge for doctors. To make the best treatment decisions, a multidisciplinary team and an individual approach to each patient should be chosen. As reviewed [7] [8], several surgical strategies are available for the treatment of PJI. These include arthrodesis, antimicrobial suppression without surgery, amputation, open or arthroscopic debridement without removal of the prostheses or resection of the prostheses with or without reimplantation. The latter can be performed either in one step (at the time of removal) or in two steps (usually delayed by weeks). All our patients were treated with this last option, a two-step surgical strategy described as the best treatment to eradicate the infection with better patient recovery [8]. To our knowledge, there is no other study that provides a treatment strategy for patients with PJI and concurrent sepsis. Since the rapid onset of sepsis treatment is known, one of our main goals was to start the first surgery as soon as possible after admission and diagnosis of the patient. This time span was 5.45 hours on average in our study group. Although this period of time seems long at first, it is due to the fact that laboratory results have to be awaited and the surgery logistics have to be adhered to. It has been shown that a longer operation time is associated with a higher risk of infections and other complications such as heart disease, pneumonia, kidney failure and sepsis [42]. Each additional 15 -20 minutes of surgery time is associated with a 9% -25% increased risk of joint infection [23] [43]. With an average of 23 minutes of surgery time in the first operation, the perioperative burden on the patient with sepsis was limited, with low blood loss.
The second planned operation took place after an average interval of 8 days and lasted 115.7 minutes on average. By this time, the patients had already improved in terms of the infection laboratory parameters and clinical symptoms, so that the long duration of the operation was associated with less risk for them. In addition, the first microbiological results were available at this point in time, which allowed for targeted antibiotic treatment.
Our results showed a high success rate of 100% in treating the infection in 10 patients who chose surgery without removal of the prostheses as a treatment option. This shows in our patient collective that the maintenance of a prosthesis in case of PJI with sepsis is justified and has a good prognosis overall. Patients who had to have their prostheses removed and re-implanted with a time delay also showed a high success rate of 100% (8 patients). These in a total follow-up period of 9.4 months on average. These confirm the validity of other studies in which the success rates of a two-stage surgical treatment were examined at 87% -100% and 72% -95% for hip and knee PJI [8]. CRP is considered and inflammatory marker for many diseases and medical conditions, however, it is supportive and not definitive evidence criteria for diagnosing PJI [8], due to its lack of specificity. One must take also into consideration that in some cases, especially in patients infected with Propionibacterium acnes CRP might be in normal levels even when every other parameters suggests an infection [45]. However, the lowering of its levels strongly suggests an effective treatment as shown in our patients, in which similar to fever, its levels started to lower after the first surgery, this after 24 hours of admission. It is known that CRP half-life is known to be around 19 hours, that its concentration peak is reported to be at 48 hours [46]. However, it has been demonstrated that after total joint replacement this parameter may have a longer half-life of 62 hours [47], which might explain the reason of some peaks on the daily measurement levels of this parameter in our patients. However, a decrease pattern can clearly be seen in our results until the last day of the inpatient care. CRP release by hepatocytes is mainly controlled by cytokines release, specially IL-6 which has been proposed as a better inflammatory and follow up marker due to its concentration peak at 6 hours and a half-life of 15 hours [47] and therefore a rapid return to its normal concentrations in comparison with CRP.
As with the majority of studies, our retrospective study design is subject to certain limitations. One possible limitation was the number of patients included in the research, which must be larger in order to draw more objective conclusions. Since PJI with sepsis as a disease is rather rare, the patient population is too diverse in terms of prosthesis design, bacterial species and their resistance patterns, secondary diseases and age, it is difficult to conduct a prospective study on the same topic. The other limitation was the short follow up time.

Conclusion
In summary, the combination of acute PJI and SIRS is a complex severe clinical picture with leading symptoms such as joint pain and feeling of illness in a pro-dromal stage before the onset of sepsis. This period of time, in our retrospective study 5 days, should be effectively used diagnostically and therapeutically to prevent the risk of developing sepsis and thus an increased mortality. Our multi-stage surgical treatment concept of a quick and above all short first operation followed by a delayed second scheduled revision operation has proven to be successful in treating PJI in sepsis. Especially the prosthesis retention attempt seems to have a good prognosis for patients in this highly acute infection situation. The CRP and fever are important progression parameters that can indicate early successful treatment. The discrepancy in the positive microbiological culture results between the early joint puncture, the subsequent first operation and above all the blood cultures from the day of admission was striking.