Effectiveness of Using Closed-suction Drainage in Treating Proximal Femur Fracture With Cephalomedullary Nail

Participants This study protocol was approved by the Institutional Review Board of the National Cheng Kung University Hospital (IRB: B-BR-107-046-T). This single center study was performed between May 2018 and April 2019. All the 60 patients who were treated with a PFNA system (Depuy Synthes®, Proximal Femoral Nail Antirotation (II), Solothurn, Switzerland) were consecutively included. Patients were included in the study if they (1) had an acute (injury less than 2 weeks) proximal femoral fracture, classified as Arbeitsgemeinschaft für Osteosynthesefragen (AO) 31A1, 31A2, and 31A3, (2) were treated with closed reduction or mini-open reduction, and (3) were followed-up for at least 2 weeks post-surgery. Patients were excluded from the study if they had (1) a history of muscular-skeletal disorder or inflammatory arthritis, (2) other combined injury, (3) a previous hip surgery, (4) an extensive open reduction of fracture site, (5) hemodynamic instability, (6) consumed non-steroid inflammatory drug, steroid, or morphine before treatment, and (7) abnormal preoperative laboratory data, especially anticoagulation data, including prothrombin time, activated partial thromboplastin time (aPTT), and platelet level. The investigators established a table of random numbers, which was readily generated by computer 15, and divided participants into the drained and undrained groups based on their arrival time to the investigator's hospital.

Surgery procedure

Experienced trauma surgeons performed all the procedures. General or spinal anesthesia was determined by anesthesiologist. All patients received a single dose of 1 g cefazolin as preoperative prophylactic antibiotic. All the operations were performed on a traction table in supine position under fluoroscopic guidance. Reduction was achieved with closed reduction or mini-open reduction, including percutaneous Kapandji procedure or a minimally invasive procedure using the cerclage wire passer (Cerclage passer, DePuy Synthes, Solothurn, Switzerland). The reduction was regarded acceptable if no significant varus malalignment of the proximal fragment (neck-shaft angle more than 15°) existed or if less than 10 mm of translation existed between the main fragments as described previously. After achieving reduction, the skin was incised 3-5 cm proximal to the tip of the greater trochanter followed by straight longitudinal incision in the fascia of the gluteus muscle. The gluteus maximus muscle was split using the blunt dissection technique to gain access to the tip of the trochanter, which was followed by the insertion of the intramedullary guide wire. A protection drill sleeve was used to the trochanter and progressive reaming was performed until the diameter was 1 mm greater than that of the selected nail. After proximal reaming up to 16 mm, nail was manually inserted up to an adequate depth under fluoroscopic guidance, which was followed by positioning of blade guide wire. The blade was positioned in the center or distal third of the femoral neck on the anteroposterior view and in the middle third of the femoral neck on the lateral view. The blade was inserted and locked with the closure of fracture gap that was followed by application of a distal screw with gig and protection sleeve. End caps were not applied in all the cases because nearly none of the implants were to be removed. In the drained group, a hemovac drain (Zimmer Biomet, Autotransfusion System [HAS], United State) was placed beneath the fascia of the gluteus muscle, surrounding the entry point of the nail at the fracture site. The fascia of the gluteus muscle was repaired with absorbable sutures, Vicryl 1-0. The duration of operation was defined as the interval between the initiation of reduction and the wound closure.

Postoperative care All the patients received a prophylactic antibiotic therapy of 1 g cefazolin every 8 hours for 1 day after surgery. All wounds were covered with a surgical dressing only and wound care was practiced on postoperative days 1, 2, and 4. The drain was removed within 24 hours based on the drainage condition, except when more than 250 ml fluid was drained in 24 hours 2, 12. Hb and Hct were examined on postoperative days 1, 2, and 4. The criterion for postoperative BT was a hemoglobin level of less than 8 mg/dL or presence of symptoms indicative of hypoxia. To control postoperative pain due to the influence of different anesthetic methods, the investigators measured visual analog score (VAS) at rest from postoperative day 1. The basic regimen after surgery was a combination of oral analgesics, acetaminophen and etoricoxib. After receiving the combination therapy, if the VAS score was more than 5, the investigators administered a subcutaneous morphine injection for pain control.

Postoperative rehabilitation All the patients were encouraged to mobilize as early as possible to prevent complications associated with immobilization. Passive rehabilitation, including passive range of motion of the affected lower limb and practice for sitting at the edge of the bed, was initiated on postoperative day 1. Partial weight bearing ambulation and rehabilitation with walker was commenced on postoperative day 2. The extent of weight bearing was adjusted by the surgeon based on the stability of the fracture and the physical strength of the patient. Morphine was administered to patients when required during the rehabilitation period.

Data collection One independent observer who was not involved with the treatment performed the clinical assessments and laboratory data collection during the preoperative and intraoperative periods and on postoperative days 1, 2, and 4 (during admission) and on postoperative day 10 for suture removal.

Primary outcome The investigators defined TBL and BT as primary outcome variables that were calculated in millimeter. TBL was measured using the Mercuriali's formula because of its accuracy, which was based on previous studies. Mercuriali's formula: TBL = blood volume (BV) × (Hct [preop] - Hct [POD 4] + transfused RBC [mL])

Secondary outcome The clinical outcomes included body temperature, VAS (for pain assessment), opioid dosage (morphine equivalent), width of the injured proximal thigh (20 cm distal to the ipsilateral anterior superior iliac spine [ASIS]), wound condition, and the other laboratory data. The wound condition was assessed based on criteria described previously with specific reference to hematoma formation and wound infection. The severity of hematoma surrounding the wound site was classified into 3 grades as follows: none, moderate (ecchymosis without swelling), and severe (ecchymosis with swelling of the wound, resulting in tension on the skin sutures) [Fig. 2]. Wound infection was classified in 2 groups: superficial infection and deep infection. A wound infection was defined as superficial when 3 of the following symptoms occurred: fever, wound discharge, and wound redness, which usually could be controlled with antibiotic prescription. A deep wound infection involved the tissue beneath the fascia of the gluteus muscle and the implant, which usually required debridement or revision. The investigators recorded hemoglobin (Hb) and hematocrit (Hct) levels on postoperative days 1, 2, and 4 as secondary laboratory outcomes.

Statistical analysis The primary objective of the study was to determine whether routine drainage in PFNA fixation for ITFs is useful. Descriptive statistics, including estimated mean and standard deviation, for continuous variables and percentages and frequencies for categorical variables were tabulated and presented. Because the sample size of both the groups was small, a nonparametric method, the Wilcoxon rank-sum test was used for comparing continuous variables and the Fisher's exact test was used to analyze the categorical variables. Box and whisker plots were used to demonstrate the BT and TBL distribution between the drained and undrained groups for intertrochanteric fracture treated with PFNA. All statistical tests were 2-sided and a p value less than 0.05 was considered statistically significant. All statistical analysis was performed with R version 3.6.1 for Windows. The investigators calculated the required sample size that would provide 90% power (α = 0.05) to the trial for detecting the difference in TBL and BT based on previous studies 12, 21. Hence, the investigators enrolled 30 patients in each group.