1 Introduction

Periprosthetic joint infection (PJI) is a serious complication of joint arthroplasty that places a significant burden on both patients and the healthcare system.^1,2 PJIs occur in 1–2 % of arthroplasties and is the primary cause for revision arthroplasty and causes substantial patient morbidity and mortality while placing significant burden on healthcare resources.¹ According to the 2021-22 Canadian Joint Replacement Registry, infection was the leading cause of hip and knee revisions at 32.5 % with revisions costing 73.0 % more and prolonging over twice the length of hospitalization compared to primary replacements.³ Given the increasing incidence of primary joint replacements, PJI's projected economic burden in 2030 is $1.85 billion in the United States.⁴ Therefore, understanding risk factors and subsequently ways to address them is prudent.

There are many known risk factors involved in PJI's,⁵ however, the effect of seasons on PJI is not as well described with controversial findings. Previous studies in Switzerland found an association between hotter temperatures in the summer months and the rates of PJIs following elective knee and hip arthroplasty.⁶ They found an increase in rate of PJI when temperature was above 20 °C, using over 116 000 procedures for their study.^6,7 However, studies in North America have found opposite findings. A separate retrospective cohort study done in a North American subtropical climate, using 3696 cases, found that there was no significance in the association between temperature, season, precipitation and/or humidity and the development of PJI's.⁸ Other studies have shown similar results, with higher rates of PJI in the winter compared to the summer months,⁹ and no differences in the rate of PJI between warm months (May to September) vs. cold months (October to April).¹⁰ A Medicare database analysis in the United States using over 1 million patients did find a significant effect on season for the development of a PJI for THA, however this significance was only found in the southern states.¹¹ Therefore, season and temperature are potentially factors for PJI in warmer geographic regions, while in colder geographic regions it has no effect on the rate of developing a PJI. The generalizability of these studies may be challenging given the nature of different climates of which they are conducted. A deeper understanding of the seasonal impact on PJIs is prudent, especially as the number of arthroplasties and global temperatures continue to rise each year.

Due to the conflicting results in the literature we conducted a retrospective chart review to determine if different seasons effect the incidence of acute and chronic PJIs. In addition we aimed to investigate if there was a correlation between PJI and age, BMI, surgeon, operating times, operating rooms, Diabetes, RA, end stage renal disease, congestive heart failure, alcohol or drug abuse, Charlson comorbidity index, surgical assist and ASA score.

2 Methods

Following Research Ethics Board approval (BIO-4321, Dec 12, 2023), we conducted a single-center, retrospective chart review of patients presenting to a tertiary center with periprosthetic joint infections of the hip, knee, ankle, shoulder, and elbow between April 2012–May 2024 as diagnosed using Musculoskeletal Infection Society (MSIS)¹² and European bone and joint infection society criteria (EBJIS).¹³ A total of 265 charts were identified.

The charts of patients who underwent primary total arthroplasties of the knee, hip, shoulder, ankle, and elbow at two tertiary hospitals from 2013 to 2023 who went on to develop a PJI were individually reviewed to document the season of primary arthroplasty alongside patient and surgical details to assess possible factors that may have led PJI development. Seasons were defined as Winter: Dec 22 – March 19; Spring: March 20 – June 21; Summer: June 22 – Sept 22; Fall: Sept 23 – Dec 21.

Patient details including height, weight, BMI, sex, age at primary total arthroplasty, and medical comorbidities such as diabetes mellitus, chronic suppressive medications (corticosteroids/DMARDS) for greater than six months, rheumatoid inflammatory conditions (rheumatoid arthritis, ankylosing spondylitis, gout), COPD, ESRD, severe renal impairment with eGFR less than 30, history of alcohol or drug abuse were recorded from operative reports or anesthesia records when available. The Charleston comorbidity index (CCI) was calculated for each reviewed chart appropriately.

Surgical details for each primary total arthroplasty were recorded from operative reports. These factors included the surgery date, length of procedure, OR room, the surgical staff and nursing staff present during the case, the anesthetic type, ASA score, use of warming blankets or equipment (bear hugger), administration of preoperative antibiotics, and administration of anticoagulation medication including the type and indication (therapeutic or prophylactic).

A total of 265 PJIs were identified. 151 of these patients were excluded based on our exclusion criteria: 1) previous PJI within the last 2 years; 2) revision arthroplasty of the affected joint; 3) patient was already on suppressive antibiotics; 4) the chart was missing critical data related to the study. This left 114 to be analyzed (See Fig. 9).

Data collection occurred in July 2024. Abstractors were blinded to each other's findings. Data was not analyzed until all charts were completed. Chart abstractors did not have a pre-existing relationship or known previous interaction with included patients and charts were randomly assigned and assessed by chart abstractors. Chart abstractors all had previous experience with paper and electronic chart reviews and received a workshop on paper and electronic chart reviews taught by the Principal Investigator.

3 Statistics

The study compared anthropometric and comorbidity data between two patient groups, those with acute infection and those with late infection. Fisher's exact test was used for categorical variables, while the t-test was applied to continuous variables. To further assess the relationship between these variables and group membership, Cramer's V was used for categorical data, and Pearson's r coefficient was used for continuous data to measure the strength of the correlations.

4 Results

We included 114 patients with a joint replacement (TKA, THA, Total shoulder-elbow-and-ankle replacement) with a subsequent PJI. The infection rate for THA at our institution for the year 2023–2024 was 1.08 % compared to 0.9 % for TKA during the same period. The average age in the acute PJI group was 64.88 years compared to 63.4 years in the late PJI group (p = 0.465). Acute THA PJI's occurred in 23/42 (54 %) cases compared to 19/42 (45 %) in acute TKA PJI's. In the late PJI, TKA occurred in 45/60 (75 %) cases compared to THA (14/60; 23 %) and total shoulder arthroplasty (1/60; 1.6 %). The majority of acute infections occurred in males (55 %; 24/43), while females were more prevalent in late PJI's (59 %; 35/59) (p = 0.161) (see Fig. 8). The average operating time for the cases that developed an acute PJI was 70.07 min compared to the patients that developed a late infection, 65.34 min (p = 0.615). SeeTable 1 for demographics.

4.2

4.2 Secondary objectives

4.2.1

4.2.1 Significant findings

We did find a "somewhat strong” correlation between the individual surgeons and their occurrence of PJI's during the study period. Please see Fig. 2. Comparing the frequency of acute or late PJI's in the different replacements reported, we did find a statistically significant difference with more THA PJI's occurring in the acute phase while more TKA occurred in the late PJI group (p = 0.002) (see Fig. 5). The majority of patients in the acute PJI had an ASA score of 3 (51 %; 21/41) or 2 (43 %; 18/41). In the late PJI's the majority of patients had an ASA score of 2 (70 %; 29/41) (p = 0.031) (see Fig. 7).

Close

Fig. 1 Assess the relationship with PJI and seasons. J0 = Winter (22 December – 19th March); J1 = Spring (20th March – 21st June); J2 = Summer (22 June – 22 September); J3 = Fall (23 September – 21 December). M0 = Acute PJI; M! = Late PJI.

Export to PPT

Close

Fig. 2 M0 – acute PJI; M1 – late PJI; Column AC are the 16 surgeons assessed.

Export to PPT

Close

Fig. 3 M0 – acute PJI; M1 – late PJI; Column AM0 – Aspirin; AM1 – Rivaroxaban; AM2 – Aspirin and Rivaroxaban; Column AM3 – other.

Export to PPT

Close

Fig. 4 M0 – acute PJI; M1 – late PJI; Column AA is the Charlson Comorbidity Index.

Export to PPT

Close

Fig. 5 M0 – acute PJI; M1 – late PJI; Column D: 1 = TKA; 2 = THA, 3 = Total shoulder replacements.

Export to PPT

Close

Fig. 6 M0 – acute PJI; M1 – late PJI; Column AG: ASA score.

Export to PPT

Close

Fig. 7 M0 – acute PJI; M1 – late PJI; Column AI – Operating rooms used for total joint arthroplasty and PJI rates.

Export to PPT

Close

Fig. 8 M0 – acute PJI; M1 – late PJI; Column N – Males = 0; Females = 1.

Export to PPT

Close

Fig. 9 Patient inclusion and exclusion flowchart.

Export to PPT

4.2.2

4.2.2 Non statistically significant findings

We did not find a correlation between acute or late PJI and BMI (p = 0.427), age (p = 0.465), operating times (p = 0.615), Diabetes (p = 0.150), rheumatoid arthritis (p = 0.499), chronic obstructive pulmonary disease (p = 0.193), end stage renal disease (p = 0.394), congestive heart failure (0.305), alcohol or drug abuse (p = 0.268), Charlson comorbidity index (p = 0.524), operating room (p = 0.621) or surgical assistant (p = 0.594) (see Fig. 7). Looking at postoperative anticoagulation (Aspirin, Rivaroxaban, Aspirin and Rivaroxaban, other) we did not find a statistically significant difference between the different groups (p = 0.314). However, we observed the highest rate of acute and late PJI's with Rivaroxaban usage (47 % and 57 % respectively). Aspirin was found to have been used in 25 % of acute PJI's and 20 % of late PJI (SeeFig. 3). We observed the majority of acute infections in patients placed on prophylactic anticoagulation (80 %; 33/41) compared to therapeutic anticoagulation (17 %; 7/41). This was a similar finding looking at late PJI's (75 % prophylactic anticoagulation). Comparing bear huggers to warming blankets we found the majority of acute and late PJI's occurred in the “warming blankets” group (80 %; 33/41). 85 % of patients received a spinal anesthetic in the acute and late PJI group. Six percent of patients had an inflammatory condition in the acute PJI compared to 10 % in the late presentation. Nine percent of patients were on chronic suppressive medications in the acute PJI group compared to 10 % in the late PJI group (steroids/biologics). The majority of patients in the acute and late PJI group had a Charlson's Comorbidity Index of 2 (30 %, 36 % respectively) SeeFig. 4.

5 Discussion

In this article we did determine that acute PJI's mainly occur in the Winter and Summer months compared to late PJI’ that occur mainly in the Fall and Winter months, albeit not statistically significant. There was a “somewhat strong” correlation between individual surgeons and frequency of PJI's. We did find more acute PJI's in THA's; compared to late PJI's mainly occurring in TKA's. Acute PJI's had an ASA score≥2 while late PJI's had an ASA score of 2. We did not find a correlation between PJI's and BMI, age, operating room, operating room nurses, Diabetes, COPD, rheumatoid Arthritis, surgical assistant and end stage renal disease, congestive heart failure. Alcohol or drug abuse and Charlson comorbidity index.

The seasonal effect on the rate of PJI's is very controversial. Many studies have been performed to try and determine if certain seasons affect the occurrence of PJI in THA or TKA's.^{6–11,14–17} Most of the studies are retrospective and include cohorts ranging from 750 patients to 1 311 672 patients.^{6–11,14–17} Some studies assessed acute postoperative infections <30 days⁹ while others included patients with PJI's occurring in the 2 years following a TKA.⁸ This heterogeneity of the studies does make interpreting the results more difficult. In our study we did not find a clear correlation between season and PJI's. This was a similar finding in other studies.^6,8,10,15,16 However, multiple studies did find a correlation between summer months and PJI's^14,9,11,17 while only one study found an increase in PJI's during the Winter months.⁷ While we did not see a clear correlation between season and the occurrence of PJI's, it highlights that other factors i.e. geographical location, medical comorbidities etc. might play an important role in the frequency of PJI's and should be included in future studies.

We did find a somewhat strong correlation between individual surgeons and PJI's. To explain these findings might be quite difficult seeing that multiple factors could contribute to these findings, for example preoperative measures i.e. hair removal with clippers prior to surgery might reduce the risk^18,19; nasal decolonization; preoperative antibiotics (type, dosage, timing); intraoperative measures i.e. operating room traffic²⁰; intraarticular irrigation solution (type, strength, duration); wound closure technique (antibiotic sutures compared to staples)¹⁹; and postoperative measures i.e. postoperative extended antibiotics for high risk patients¹⁹ to name a few. Surgeons use different techniques and therefore it is imperative for all surgeons to familiarize themselves with up-to-date research to optimize patient care.

We did find a strong association between acute PJI's and an ASA≥2. Multiple studies confirmed these results with an ASA score≥2 to be a risk factor for PJI's in THA and TKA's.^{21–23,24,18} Interestingly some studies only identified the association between ASA score and PJI in THA patients.^25,26,27

We did not find a clear correlation with obesity (BMI >30kg/m2) and increased risk of PJI. Multiple systematic reviews have been performed with conflicting results. Two large systematic reviews evaluating obesity and their role in PJI's concluded that PJI risk increased with obesity. The first study included 581 012 THA patients while the second systematic review included 917 447 TKA patients.^28,29 Similar findings were recorded in patients that failed a two-stage revision.^30,31 Interestingly this was not demonstrated in a systematic review evaluating obesity and unicompartmental knee replacements. In this review they included 80 798 UKA patients and did not see an increased risk of PJI with obesity.³² Even though multiple studies did identify an association between PJI's and older age,^25,27,18 some proved the opposite.²³ We did not find a correlation between age and increased risk for PJI's. Similarly, we were unable to find a correlation between sex and PJI's. This was a similar finding in other studies.^26,24 The studies that did find a correlation was inconclusive, with some studies suggesting males are an increased risk^23,27 while others cite female gender to be an increased risk.²⁵ This was similar finding for operative duration. Multiple studies concluded that increased operative time increased PJI risk,^25,23,24 while others demonstrated no correlation.^26,27 Potential reasons why we could not demonstrate a correlation could include a smaller sample size, heterogenous group or confounding factors.

We did find the highest rate of acute and late PJI's with patients on Rivaroxaban for postoperative prophylaxis. One explanation can be that Rivaroxaban was the treatment of choice for postoperative prophylaxis following a joint replacement. Recently our center has adopted Aspirin usage for postoperative prophylaxis. Evaluating the literature the majority of studies do show non-inferiority with ASA and deep venous thrombosis or pulmonary embolus prevention.^33,34,35,36 Most studies also did not see a statistical difference in infection rate and mode of postoperative DVT prophylaxis.^33,37 One study looking at PJI and different postoperative anticoagulation found that ASA and/or compressive devices had a decreased risk of bleeding and PJI's in revision cases³⁵

6 Conclusion

This study examined periprosthetic joint infections (PJI) following THA, TKA replacements. Acute PJIs were more common in THA, while late PJIs were more frequent in TKA. No significant correlations were found with factors like BMI, age, or comorbidities, but there was a notable association between individual surgeons and PJI rates. Although not statistically significant, Rivaroxaban use showed higher PJI rates. Seasonal variations in infection rates were minimal, with slightly higher rates in winter. These findings, while insightful, are limited by the study's small sample size and single-center design.

CRediT authorship contribution statement

Mars Yixing Zhao: Methodology, Formal analysis, Investigation, Writing – original draft, Writing – review & editing, Supervision, Project administration. Evan Parchomchuk: Investigation, Resources. Thomas Goldade: Investigation, Resources. Mikayla Rudniski: Investigation, Resources. Nathan Oster: Investigation, Resources. Mason Beaulieu: Investigation, Resources. Michaela Nickol: Writing – review & editing, Supervision, Project administration. Johannes M. van der Merwe: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration.

Informed consent

N/A.

Patient and guardian consent

This is not applicable to this study – No identifiable names, images, videos are in the manuscript.

Ethical approval and consent to participate

Approved on the December 12, 2023; Bio 4321 NER.

Availability of supporting data

All the raw data and materials described in the manuscript is available upon requests to any scientist wishing to use them for non-commercial purposes.

Consent for publication

Not applicable.

Ethics statement

Research Ethics Board approval (BIO 3704, 12-Dec 2023).

Data available statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.