Low post-arthroplasty Infection rate is possible in developing countries Long-term Experience of local Vancomycin use from a high-volume tertiary Knee center in Iran

Background: Utilizing intrawound vancomycin powder in TKA surgery has yielded rather contrasting results in the current literature. Furthermore, CDC criteria, although effective in general, are not speci�cally designed for post-TKA infections. Here, we present a 7-year experience of vancomycin use in primary TKA in a high-volume tertiary knee center in Iran. Also, new criteria are proposed to detect suspected super�cial post-TKA infections. Methods: This is a retrospective analysis of primary total knee arthroplasties performed in a tertiary knee center, from March 2007 to December 2018, by a single senior knee surgeon. All patients with follow-up periods of less than one year were excluded from the study. Since March 2011, all patient received vancomycin (powder, 1gr) before water-tight closure of the joint capsule. Comparison was made between this group and historical control subjects (operated from March 2007 to March 2011). Results: Altogether, 2024 patients were included in the study. The vancomycin and the control group included 1710 and 314 cases respectively. Patients were mostly women (male to female ratio: 1 to 4), with a mean age of 65.20 (SD=10.83) years. In the vancomycin group, the rate of suspected SII (1.87%) and PJI (0.41%) was signi�cantly lower than the control group (P= 0.002). Conclusions: Our experience shows that by utilizing intrawound vancomycin as a routine practice along with other measures, we were able to reach relatively low rates of deep post-TKA infections. Howbeit, randomized controlled trials are required to clarify the effect of intrawound vancomycin on post-TKA infection rate.


Background
Joint replacement enhances function and alleviates pain in a completely destructed knee, but requires implementing high technical perfection by surgeons or it could result in devastating consequences (1).Deep periprosthetic joint infection (PJI) following total knee arthroplasty (TKA) usually needs treatment by two-stage revision arthroplasty (2,3).It causes signi cant bone loss, as well as substantial injury to soft tissue structures and could necessitate higher-constraint prostheses with lower reported survivals (4); therefore, even the lowest infection rate is too high to be accepted by the surgeon or the patient after TKA and reducing infection rate has always been a challenge.This is especially true in Iran, where heavy economic sanctions have caused revision devices and prostheses less available (3)(4)(5)(6)(7).Thus, PJI after TKA, is an intolerable complication in Iran.
Much research has been carried out for prevention and also early detection of post-TKA infections (2,5,6,(8)(9)(10)(11).Regarding prevention, a recent topic is utilizing intrawound vancomycin powder.Separate in vivo rat investigations (12,13) have demonstrated the effectiveness of intrawound antibiotics in clearing Staphylococcus aureus from contaminated femoral implants.Intraoperative use of vancomycin powder has also been shown to be effective in reducing the chance of infection in spine surgeries (14)(15)(16)), but literature on its use in TKA surgery has yielded rather contrasting results (17)(18)(19).
Regarding early detection, the latest consensus on PJI has provided some valuable criteria to diagnose a deep infection (20), but literature on the detection of super cial infection after TKA is not clear (21,22).Some studies (22) have used the US Centers for Disease Control and prevention (CDC) criteria for identifying surgical site infections which were de ned by the CDC in 1992 as those occurring within 30 days of surgery (23).CDC criteria although effective in general, are not speci cally designed for post-TKA infections.In fact, most TKA patients experience pain, tenderness and swelling at the surgical site.Furthermore, it does not seem logical to wait for wound drainage and positive cultures to intervene.
In the current study super cial incisional infections (SII) were suspected, in case of any signs of erythema and/or warmth and/or itching and/or increased local pain on surgical site.We present a 7-year experience of vancomycin use in a high-volume tertiary knee center in Iran and we set out to report our ndings as well as discuss the reasonable explanations.

Methods
This is a retrospective analysis of primary total knee arthroplasties performed in a tertiary knee center, from March 2007 to December 2018.Intrawound vancomycin was used since March 2011 in all TKAs.
The arthroplasties performed prior to 2011 were considered as the historical control group.Other infection prevention protocols and surgical technique were the same between the two groups.All arthroplasties were performed by a single senior knee surgeon.Cases with follow-up periods of less than one year were excluded.Records of each patient's age, gender, operation time and length of stay in hospital after surgery were collected.Rates of super cial and deep infection are reported along with a detailed report on comorbidities of the infected cases.
We classi ed all patients with any of the following 4 criteria including erythema and/or warmth and/or itching and/or increased local pain at the surgical wound as suspected cases of SII and treated them with a one-week course of oral Levo oxacin 500mg twice daily.None of them were proved to be infections, as no culture sample was analyzed.In this retrospective analysis, all PJI cases met the criteria proposed by the proceedings of the Philadelphia consensus (20).
All patients underwent cemented TKAs without patellar replacement (NexGen, Zimmer Biomet prostheses until July 2017 and Persona Zimmer Biomet prostheses since then).As an infection prevention measure since March 2011, intrawound vancomycin (powder, 1 gr) was used at the end of TKA surgeries right before tightly closing the joint capsule.It is worth mentioning that in addition to intrawound vancomycin, intra-articular injection of tranexamic acid was also performed as a routine practice in all cases since 2007, and its use was previously published from this center (24).Data are reported as means, standard deviations (SD) and percentages.Comparing percentages was performed by the StatPac software (StatPac Inc. 2017).Two sample t-test was used to compare means.Statistical analysis was performed using SPSS software version 25 (IBM corporation).Signi cance was de ned as a P-value under 0.05.

Results
Amidst a total of 2494 TKAs performed during the study time period, 2024 patients had a minimal documented follow-up period of one year and were included in the study.Demographic data are depicted in Table1.Patients were mostly women (male to female ratio: 1 to 4), with a mean age of 65.20 (SD=10.83)years.Demographic characteristics were the same in the control (n=314) and the vancomycin groups (n=1710).
Control group: among the 36 patients (11.46%) who showed signs of suspected super cial incisional infection and were treated with one weeks of oral Levo oxacin, none lead to deep infection.Meanwhile, deep infection occurred in 6 patients (1.91%).One of them was early PJI (3 weeks after surgery) and underwent irrigation and polyethylene exchange.The other 5 were late PJI and underwent 2-stage revision surgery.
Vancomycin group: Taking both suspected and de nite infections into account, the overall infection rate was 2.16% (37 of 1710 patients), including 7 periprosthetic deep joint infections and 32 suspected SIIs.Besides, from the 32 suspected super cial infections that were treated with oral Antibiotics, all reported disappearance of symptoms yet 2 cases returned with late PJI.Except 7 cases, all cases of suspected SII were detected close to one month after surgery.A detailed report on the comorbidities of the patients who developed super cial infection, is provided in Table 2.The most common characteristics among this group of patients were age > 60 years, followed by obesity and diabetes mellitus in order of prevalence.Except one, every one of them had at least one comorbidity.All PJI cases were females.There were no cases of early PJIs de ned as those occurring over the rst 6 weeks after surgery.Thereupon, no polyethylene exchange was performed over the studied time frame.All in all, 7 patients (0.41%) underwent two-stage knee arthroplasty due to deep joint infections.
Comparison: Regarding baseline characteristics, there were no signi cant difference among the groups (Table 1).In the historical control subjects, the rate of suspected SII (11.46%) and PJI (1.91%) was signi cantly higher than the vancomycin group (P= 0.002).

Discussion
This is a retrospective analysis of an 11-year experience of a single knee surgeon from a referral knee center and a total of 2024 consecutive primary TKAs were enrolled.Patients include those from all over Iran.Intrawound vancomycin powder was used in 1710 cases and the results were compared with a historical control group with 314 cases.The mean age, mean duration of surgery and mean Length of stay along with data on comorbidities and nal outcome of those who were diagnosed to have super cial or deep infections are presented.
The mean length of stay in our study was 3.21 days (SD=2.05),which is somewhat higher than recent reports (25,26).In our center, preoperative planning and paraclinical work-ups of patients were performed on an outpatient basis.Only after determining precise surgical plans, having prepared required devices, and optimizing their medical states, were TKA candidates scheduled for surgery.Patients were admitted to the hospital early in the morning and then underwent surgery the same day and if alert and well, patients were discharged the next morning.The only exceptions were those patients coming from distant cities or rural places of the country who neither had a place to stay in Tehran nor could come back any time sooner than 2 months after surgery.Since a strong family physician-based postop care system is not available in Iran, this small group of patients were discharged after 1 to 2 weeks, until knee range of motion was secured and the surgical wound completely healed.Of our TKAs, 90% were discharged in less than 6 days and 60% were discharged in less than 3 days after surgery.
Mean age in our study population was 65.20 (SD=10.83),which is somewhat younger than that of similar studies (22), despite the fact that people here refer for surgery at the latest stages of knee joint destruction.This obvious lower age of TKA candidates in Iran could be due to the Middle-eastern lifestyle which is characterized by high-exion activities such as praying, sitting on the ground and using Iranian toilets which need squatting.
In the vancomycin group, there were 32 cases (1.87%) of suspected super cial incisional infection and no cases of early PJI.Seven cases of late PJI (0.41%) were diagnosed and underwent two-stage revision arthroplasties, among whom 2 patients had a history of being treated for suspected SII.Our rate of PJI is in concert with recent reported rates (22).
A discussion needs to be made on how to diagnose super cial incisional infections after TKA.To the best of our knowledge, there is no uni ed approach in the literature on this matter and CDC criteria are not designed, speci cally for post-TKA infections.Our approach was to detect all cases suspected of having a super cial incisional infection and treating them with a short course of antibiotics.We can't be sure if these 32 cases were in fact, infections, or solely in ammations or hypersensitivity reactions to Monocryl stitches.There was a 6-month period of time from April 2016 to September 2016, when an increase in these cases was observed (10 cases).After intense scrutiny, we found similar reports from other centers in Tehran using the same brand of Monocryl as ours.After switching to rapid Vicryl, the SII rate declined.The role of stitches in TKA wound complications has been reported before (34).We cannot still be sure how many suspected SIIs were actually stitch reactions.Howbeit, the fact that most cases happened close to one month after surgery (the time that Vicryl starts to resorb) makes this explanation more convincing.Indeed, it seems that even the quality of stitches we used has been seriously affected by tight economic sanctions.There were no records of SII over the last year of the study time.
There were some key routine practices that possibly prevented higher infection rates in our study.
First and foremost, it is the oriented and highly trained operating room (OR) personnel who understood the importance of infection control and strict infection prevention standards.Second is our approach to early wound infection.During postop follow-up examinations, if the surgical wound showed signs of erythema and/or warmth and/or itching and/or increased local pain at the surgical wound, super cial incisional infection was suspected and a one-week course of oral Levo oxacin 500mg twice daily was prescribed.This might have been a factor in the low rate of PJI among our recorded cases of SIIs.We do not use CDC criteria for surgical site infection, as it is not speci cally designed for knee arthroplasty.Pain, tenderness and/or swelling which are mentioned in CDC criteria, occur in the early postoperative period of most TKAs and do not seem speci c enough to diagnose a super cial infection.In fact, in knee arthroplasty, due to cumbersome management and poor outcomes of PJI, we need early diagnosis of any super cial infection and wound drainage or positive culture mentioned in the CDC criteria would procrastinate diagnosis and prevent early intervention.The third probable factor, could be the use of intrawound vancomycin powder before water-tight closure of the joint capsule.Since we started routine use of vancomycin powder in the wound at the end of TKA surgeries the infection rate decreased signi cantly (P=0.002).The effect of vancomycin powder on infection rate has not been proved in TKA (17)(18)(19), in contrast to spine surgery where it has been well established (14)(15)(16)).Although we report promising results, its effectiveness cannot be proved based on the current study either.
The strengths of this study include a large sample size and enrolling consecutive patients to decrease selection bias.Also, all the arthroplasties were performed by a single surgeon with a uni ed surgical technique, which eliminates the confounding effect of different surgical techniques.Nevertheless, there is an inherent probability of selection bias due to implementing historical control subjects and lack of randomization (35).Despite our results, determining the exact effect of intrawound vancomycin on the rate of deep infection after TKA, needs prospective randomized controlled trials designed speci c for its use in knee arthroplasty surgery (36, 37).

Conclusion
Our experience shows that by utilizing intrawound Vancomycin for 7 years as a routine practice along with other measures, we were able to reach relatively low rates of super cial and deep post-TKA infections.