Incidence and predictors of surgical site infection after distal femur fractures treated by open reduction and internal fixation: a prospective single‐center study

Background There remain limited data on the epidemiological characteristics and related predictors of surgical site infection (SSI) after open reduction and internal fixation (ORIF) for distal femur fractures (DFFs). We designed this single-centre prospective study to explore and forecast these clinical problems. Methods From October 2014 to December 2018, 364 patients with DFFs were treated with ORIF and followed for complete data within one year. Receiver operating characteristic (ROC) analyses, univariate Chi-square analyses, and multiple logistic regression analyses were used to screen the adjusted predictors of SSI. Results The incidence of SSI was 6.0 % (22/364): 2.4 % (9/364) for superficial SSIs and 3.6 % (13/364) for deep SSIs. Staphylococcus aureus (methicillin-resistant S. aureus in 2 cases) was the most common pathogenic bacteria (36.8 %,7/19). In multivariate analysis, parameters independently associated with SSI were: Open fracture (OR: 7.3, p = 0.003), drain use (OR: 4.1, p = 0.037), and incision cleanliness (OR: 3.5, p = 0.002). An albumin/globulin (A/G) level ≥ 1.35 (OR: 0.2, p = 0.042) was an adjusted protective factor for SSI. Conclusions The SSI after ORIF affected approximately one in 15 patients with DFFs. The open fracture, drain use, high grade of intraoperative incision cleanliness, and preoperative A/G levels lower than 1.35 were significantly related to increasing the risk of post-operative SSI after DFFs. We recommended that more attentions should be paid to these risk factors during hospitalization. Trial registration NO 2014-015-1, October /15/2014, prospectively registered. We registered our trial prospectively in October 15, 2014 before the first participant was enrolled. This study protocol was conducted according to the Declaration of Helsinki and approved by the Institutional Review Board. The ethics committee approved the Surgical Site Infection in Orthopaedic Surgery (NO 2014-015-1). Data used in this study were obtained from the patients who underwent orthopaedic surgeries between October 2014 to December 2018.

(Continued from previous page) this study were obtained from the patients who underwent orthopaedic surgeries between October 2014 to December 2018.
Keywords: Predictor, Surgical site infection, Distal femoral fracture, Internal fixation surgery Background Distal femur fractures (DFFs) are relatively uncommon but severe in orthopaedic trauma, comprising approximately 8.7 % of all femoral fractures and 0.8 % of total body fractures in Chinese adults [1]. These fractures show a bimodal distribution. On the one hand, most of the fractures in younger patients result from high-energy injuries, which are usually open and comminuted fractures [2,3]. On the other hand, fractures in older patients are caused by low-energy injuries, with a one-year mortality rate up to 13.4 % [4]. Furthermore, 50 % of the DFFs potentially involve articular surface [1], and operative intervention is the most common option for these patients such as open reduction and internal fixation (ORIF) or close reduction and internal fixation (CRIF). However, excessive soft tissue dissection during operation may impair the already damaged soft tissue, possibly leading to postoperative complications such as surgical incision site, knee dysfunction and bone union [5,6].
Of these complications, surgical site infection (SSI) is a major challenge for most orthopaedic and trauma surgeons. SSIs are known to increase the length of stay by an average of 10 days and to cost the National Health Service (NHS) of the United Kingdom an estimated £700 million per year [7]. It has been reported that about 50 % of SSI cases can be avoided by the application of evidence-based prevention strategies [8]. Therefore, the identification of SSI-related predictors and screening of at-risk patients may propose a cost-effective and simple approach for prevention of SSI occurrence. Recently, most researches concerned SSIs of the hip, tibial plateau, and ankle [9][10][11][12][13].
However, studies on the epidemiological characteristics of SSIs after adult DFFs treated by ORIF were rare. The incidence of SSI after operations for DFFs has been showed to vary from 0.3 to 11.2 % [14], and most factors are not conclusive. Hoffmann et al. identified open injury and current smoking as associated risk factors [6]. Lu et al. added obesity and diabetes mellitus to the SSIrelated predictors [15]. Moreover, all three studies above were retrospective and were collectively limited by untimely data selection owing to recall and response bias.
We conducted the prospective study for two objectives: first, to restudy the epidemiological characters of SSI after ORIF for adult DFFs; second, to recognize the SSI-related factors and their cutoff values. We hypothesized that the open fracture, cigarette consumption, and obesity were associated with increased DFF infection.

Study design
A prospective study was designed, and data were collected from patients who underwent ORIF for DFFs from October 2014 to December 2018. Administrative permissions (NO 2014-015-1) were acquired by our team to access the clinical patient data used in our research, which was granted by the Committee on Ethics and the Institutional Review Board of the Third Hospital of Hebei Medical University. All patients who were 18 years old and over with acute DFFs treated by ORIF were included in this study. The exclusion criteria were listed as following: age less than 18 years old, old fractures (> 21 days from earliest trauma), pathological fractures, and first treatment at other hospitals. Patients with open fractures or with multiple fractures were also involved to investigate their effect on SSIs. Preoperative infections of the open fractures were excluded, and the old fractures were also excluded for the different operative procedures such as bone grafting procedures. All enrolled patients had intraoperative antibiotic prophylaxis and were separated into two groups based on the occurrence of SSI. The case group was defined as patients with SSIs, and the control group involved patients who did not suffer from SSIs. The endpoint of the prospective study was any evidence of SSI obtained from telephone assessment, interview, or the clinical information one year after surgery.

SSI definition
The diagnostic criteria for SSI within one year postoperatively were based on the definition of the CDC [8]. A superficial SSI only infiltrates the skin or subcutaneous tissue of the operation site. A deep SSI was identified if it satisfied one of the following conditions: infection through the deep fascia, persistent wound effusion or dehiscence, local abscess requiring focal debridement and implant replace or retrieve. All the wound exudates from the patients were collected with swabs and sent for causative agent culture and sensitivity. Any inpatient who commenced antibiotic treatment for wound problems (redness, swelling, hot pain) but did not conform the criteria of the deep SSI was classified as a superficial SSI, regardless of any microbiology results.

Data collection and definition of variables of interest
All the data mentioned below were collected by five well-trained investigators. Investigators followed the patients closely by morning work rounds and reviewed patients' clinical data. The suture site was observed by researchers starting from the first day after ORIF until hospital discharge. After discharge, patients who had not developed SSI were followed by telephone at 2, 4, 6 and 12 months after discharge. Patients with suspicious SSI were required to return for re-examination and etiologic diagnosis. We usually recalled the patients for the radiological and clinical periodic evaluation every half a year after hospital discharge. During the study period, detailed variables of interest were collected and divided into four aspects.
Fracture-related variables included injury type (closed, open), concurrent fracture sites (single fracture, multiple fractures), affected side,and injury mechanism. Injury mechanisms were divided into two groups: low-energy (fall from a standing height) and high-energy (falling accidents from high places, human violence and others).
Operation-related variables included history of previous operation at any site, preoperative stay, postoperative stay, intraoperative blood loss (< 400 and ≥ 400 mL), operation duration (< 120, 120 to 180, and > 180 min), anaesthetic type (local, combined spinal-epidural, and general), internal fixator use (plate or no plate), intraoperative drainage use, incision cleanliness (I, clean; II, potentially contaminated; and III, contaminated), and the American Society of Anesthesiologists (ASA, I-II and III-IV) classification system [9]. The incision cleanliness was evaluated before the operation, and the dirty/infected incision was excluded from the study. Preoperative stay was defined as the time period from the first injury to ORIF, which was separated into two groups: 1, < 7 days and 2, ≥ 7 days.

Statistical analysis
Statistical analysis was conducted with SPSS version 25.0 (IBM Corp., Armonk, NY, USA). The continuous variables were showed as the median, mean ± standard deviation (SD). The distributions of all data were evaluated for normality by the Shapiro-Wilk test. A Whitney U test or t test was used to compare continuous variables between SSI and non-SSI groups according to the Shapiro-Wilk test. For the continuous variables with statistical significance (p < 0.05), receiver operating characteristic (ROC) analyses were carried out to identify the optimum cut-off value. Subsequently, categorical variables were compared by the chi-square test. Predictors with significance (p < 0.05) from the single factor analysis were entered into multiple logistic regression analyses (backward LR). In addition, surgical duration, diabetes mellitus, hypertension, cardiovascular diseases and BMI were selected into the model for the variables of interest. The odds ratio (OR) and 95 % confidence interval (95 % CI) were conducted to evaluate the correlation magnitude between factors and SSI risk. Normally a p < 0.05 was considered statistically significant.

Results
The selection of the patients Figure 1 showed the flow chart for the selection of participants. A total of 461 patients with DFFs were admitted for ORIF, and 47 patients were less than 18 years old; 19 suffered from pathological fractures; 5 had old fractures; 2 had incomplete data, and 24 were lost (6.2 %) to follow-up. 364 patients were included for the final analysis.
Diagnostic time points and incidence of postoperative SSI As shown in Fig. 2, the median time for diagnosis of SSI was 14 days after ORIF with a range from 2 to 106 days. The most of SSIs (81.8 %,18/22) was found during the subsequent hospitalizations. The total incidence of SSIs was 6.0 % (22/364). The superficial SSIs accounted for 2.4 % (9/364) and deep SSIs 3.6 % (13/364).   Patient demographic data and fracture characteristics

Frequency of causative bacteria
As shown in  Laboratory variables and the optimum cut-off value Table 4 depicted the univariate analysis of laboratoryrelated variables. A significant difference was observed for the variables of ALB level (p = 0.017) and A/G level (p < 0.001) between the two groups. Table 5 showed that ROC analysis was performed to identify the area under the curve and the optimum cut-off value for each statistically significant variable listed above. The cut-off values for ALB and A/G levels were 30.3 g/L and 1.35, respectively. Based on these cut-off values, we dichotomised the variables.

Discussion
Our present study of 364 patients indicated that the total incidence of SSI after ORIF for DFFs was 6.0 % with one-year follow-up, which was consistent with a retrospective multicentre analysis of 724 patients with a twoyear follow-up [15]. We confirmed that S. aureus was Open fracture is a well-recognized risk factor for SSIs after orthopaedic surgeries [16]. In the present study, the    [17]. All the open fractures and polytrauma patients were firstly conducted by aggressive debridement in our centre. Then, temporary traction bows or external fixators were performed according to the Gustilo-Anderson classification of soft tissue conditions at fracture sites and the time from trauma to surgical debridement [18]. Emergent one-stage procedures were performed for the acute Gustilo grade I and II fractures. Regarding Gustilo grade III fractures or Gustilo grade II fractures with the time from trauma to surgical debridement over 8 h, temporary traction bows or external fixators were performed for two-stage procedures. Drains are used extensively in orthopaedics with the purpose of reducing the postoperative seroma. However, the criteria for using drains are not clear; patients often complain of anxious and pain from drainages, and drainage sites may retain a potential infection source. Some studies indicate that drain use plays a critical role in developing SSI after surgery [19,20], which is in accordance with our results that an independent factor of drain use increased probability of SSI by 4.1 times (95 % CI, 1.1-15.5). Theoretically, drain use can increase the risk of infection. Bacteria, especially skin microbiome, spread along the drainage tube and have been identified from the tips of drainage tubes even as early as 48 h post-operatively [21]. Pennington et al. reported that long-term surgical drain retention was correlated with the risk of deep SSI after operations for degenerative spinal diseases [22].
It is well known that dirty surgical incisions have prolonged adverse impacts on wound closure. In the present investigation, we evaluated incision cleanliness during the operation. Then, we concluded a similar result that the risk of SSI increased by 3.5 (95 % CI, 1.6-7.7) times with every increase in the grade of incision cleanliness. In a retrospective case-control study that included 2617 cases of ankle fractures treated by ORIF, the OR of SSI was 1.8 (95 % CI, 1.1-3.2) with grade II-IV incision cleanliness, which is also in agreement with other reports of orthopaedic procedures [13,23]. In clinical practice, medical staff examine surgical incisions and worry about infection in the early days after surgery [24]. However, patient-directed active surgical incision self-monitoring may help to further SSI reduction [25]. Hence, this finding would be a significant advance for SSI management that integrates reliable patients' surgical incision surveillance into the clinical workflow. The higher A/G level instead of ALB level was a significant independent protective factor for SSI after adult DFFs treated by ORIF, which was first reported in the present study, although a lower ALB level had been reported as a risk factor for SSIs after traumatic and elective orthopaedic surgeries [16]. Moreover, we further found that patients with A/G levels ≥ 1.35 had a 77.0 % decreased possibility of SSI when compared to those with A/G levels < 1.35. In the clinic, the A/G level (protein quotient) is the weight ratio of albumin to globulin, which is normal range from 1.2 to 2.4. The A/G maintains a lower level in protein deficiency or metabolic abnormalities. We proposed that the A/G level was more comprehensive and meaningful for predicting SSI risk than the ALB level. Both ALB and GLOB are necessary to maintain nutrition and immune balance for wound healing. On the one hand, ALB transports essential  Abbreviation: ROC receiver operating characteristic, ALB albumin, A/G albumin/globulin electrolyte and amino acids to improve the wound healing. On the other hand, GLOB, especially immunoglobulin, has indispensable and favourable anti-infection effects as well as lowers the high risk of postsurgical infection in the intensive care unit (ICU) [26][27][28]. The present study had three highlights: first, it was a prospective study with a one-year follow-up; second, ROC analysis was performed to detect a highly sensitive cut-off value for statistically significant continuous variables; and third, to our knowledge, it was first study to report that an A/G level ≥ 1.35 is an independent protective factor for SSI after adult DFFs treated by ORIF. However, the study was not without limitations. The old fracture might be an important variable, which were relatively uncommon (only 5 patients with old distal femur fractures were excluded) in our study. Temporary fixation of open fractures was usually accomplished with traction bows or external fixators in our center. However, we did not include the variable of the preoperative external fixation use. The interference of ORIF performed by multiple trauma surgeons was not excluded. In addition, some variables that potentially influence the development of SSI were not included, such as the internal fixation material (titanium or stainless) and surgical incision length.

Conclusions
In summary, the overall incidence of SSIs after adult DFFs treated by ORIF was 6.0 % (22/364), with an incidence of superficial SSIs of 2.4 % (9/364) and of deep SSIs of 3.6 % (13/364). The open fracture, drain use, high grade of intraoperative incision cleanliness, and preoperative A/G levels lower than 1.35 were significantly associated with increasing the risk of SSI for DFFs after ORIF. We recommend careful assessment of these risk factors during hospitalization.