Consecutive cases of TKA who used portable NPWT (PICO®, Smith Nephew) immediately after wound closure in the operating room and for 7 consecutive days were prospectively evaluated from January 2016 to December 2017 (Fig. 1). The inclusion criteria were patients of any age with primary or secondary knee osteoarthritis who underwent elective unilateral arthroplasty. The exclusion criteria were patients with previous knee, femur or tibia surgeries on the side of the arthroplasty, previous osteomyelitis in the femur or tibia ipsilateral to the operated knee and patients who required the use of revision implants in the tibia or femur due to severe deformity or previous severe ligament instability. In addition, patients who were unable to perform the weekly postoperative evaluation in person in accordance with the established schedule were excluded.
Surgery was performed with the use of a pneumatic tourniquet through an anterior knee incision and trans quadricipital and medial parapatellar approach. The procedures were always performed with resection of the posterior cruciate ligament (PCL). All cases were performed with cemented components without antibiotics. The subcutaneous tissue was closed with 3 − 0 Vicryl sutures, and the skin was closed with simple 4 − 0 nylon sutures. Antibiotic prophylaxis was performed with cefuroxime for 24 hours for all cases. Drains were used for a maximum period of 24 hours, regardless of the draining volume, and the sutures were removed between the 14th and 21st days after surgery. Post-operative rehabilitation was not changed due the NWPT device, as its application does not hinder patient’s functional exercises. Range of motion and weight-bearing as tolerated were allowed since the first day after surgery.
This group was compared with a control group of consecutive cases (surgeries performed from January 2013 to December 2015) that underwent the same type of intervention using the same technique but who used conventional dressings; the first dressing was changed after 7 days, except in cases of saturation due to bleeding from the wound.
All surgical wound healing complications were evaluated weekly in the first 6 weeks. After that, the cases that progressed satisfactorily were evaluated at 3 months, 6 months and 12 months to assess eventual infections. The cases that progressed with problems or complications related to the surgical wound were evaluated weekly until resolution of the condition. The patients who used the PICO® device were evaluated when it stopped functioning after 7 days, unless there was a prior need due to saturation of the dressing.
In addition to epidemiological data and comorbidities, the following parameters related to the surgical wound were evaluated: presence of a haematoma around the wound or operated knee; persistent drainage from the surgical wound (considered as drainage for more than 7 days); hyperaemia around the surgical wound, either focal or widespread, defined as any redness around the skin incision, skin necrosis (Fig. 2), independent of requiring therapeutic treatment; dehiscence of the surgical wound (Fig. 3), superficial or deep, independent of requiring therapeutic treatment; presence of blisters around the surgical wound (Fig. 4); presence of postoperative infection (Fig. 5); length of hospital stay; and need for further intervention for any reason. Reintervention was considered as the need to take the patient to the operating room for any reason related to the arthroplasty. The incidence of deep venous thrombosis (DVT) was also recorded.
The presence or absence of all parameters related to the wound were performed by two evaluators dichotomously (yes or no). The more experienced examiner served as the main examiner, and the less experienced examiner enabled evaluation of interrater agreement.
Data And Statistical Analyses
Separate univariate and multivariate analyses were performed for the full cohort of patients.
For the univariate analysis, Pearson’s chi-square test and the Fischer test for categorical variables were performed. For continuous variables, Shapiro-Wilk tests and histogram analysis were used to check the normality of the data, and the independent samples t test or Wilcoxon rank-sum test was used for continuous variables as appropriate.
We used multivariate analysis with logistic regression to assess the independent risk factors for the presence of any complication, including infection, wound dehiscence, wound edge necrosis, blisters or persistent wound drainage.
No a priori power calculation was performed, as all available eligible subjects were included. A post hoc power calculation revealed an achieved power of 91.2% for the comparison of any complication between all patients using NPWT and the control group. For the logistic regression, the practical rule that there should be at least 10 events per variable was considered, and therefore, the sample was adequate.
The interrater agreement between the 2 examiners for the wound parameters was evaluated using the Cohen kappa coefficient.
All normally distributed continuous variables are reported as the mean +/- standard deviation, and non-normally distributed variables are presented as the median (interquartile range). Statistical significance was considered when the p value was less than 0.05. The statistical software SPSS 22 (IBM Corp., NY, USA) and G*Power 3.1.9.3 (Erdfelder, Faul, & Buchner, Universität Düsseldorf, Düsseldorf, Germany 2009) were used to perform the analysis.