A prospective, randomized, single-blinded controlled trial (ClinicalTrials.gov identifier: NCT04136431) was conducted at a single, academic, teaching, and medical hospital. After approval by the Institutional Review Board (IRB Number: 201102015B0), this study started to enroll the participants after obtaining written informed consent. All patients were enrolled by the Consolidated Standards of Reporting Trials.
Participants and setting
Patients undergoing primary CAS-TKA were assessed for eligibility from January 2017 to July 2017. Patients between 18 and 90 years of age who underwent primary, unilateral navigation-assisted TKA were included in this study. The exclusion criteria were as follows: (1) patients who underwent bilateral TKAs, unicompartmental TKA or revision TKA (2) patients who had to remove previous implants or history of high-tibial or distal femoral corrective osteotomy (3) patients who were unable to response the questionnaires.
A total of sixty-eight patients who underwent CAS-TKA were enrolled. Among them, seven patients declined to participate. One patient was excluded due to consciousness disturbance after returning the ward and unable to answer questions listed in the questionnaires. Therefore, 60 participants were randomly allocated to the intervention group (n = 30) and the control group (n = 30). All patients completed the analysis before discharge (Fig. 1).
The randomization was performed by the research staff using a parallel, 1 to 1 allocation method. A computer-generated randomization schedule was generated to assign participants to the intervention group by using a block size of 8 (1:1 ratio). The randomization occurred on the day of surgery using an opaque, sealed envelope, pre-labeled method.
All patients received a unilateral primary CAS-TKA by a single experienced surgeon. A pneumatic tourniquet was inflated to 300 mmHg pressure before the incision and deflated at the end of surgery after skin closure. The navigation system was Vector Vision (BrainLab, Heimstetten, Germany) and all TKAs were cemented using the same prosthesis (NexGen Legacy posterior-stabilized prosthesis; Zimmer Inc., Warsaw, IN, USA). There was no local infiltration of the local anesthetic in the knee joint. A suction drain was inserted before wound closure and removed on postoperative day (POD) 1. All surgical wounds, including the stab incisions for the insertion of temporary pins, were closed with interrupted skin stitches. All patients received aspirin as venous thromboembolism prophylaxis for 14 days. Multimodal pain management (acetaminophen, cyclooxygenase-2 inhibitors drugs, and tramadol/acetaminophen combination tablets [Ultracet®] were applied for all patients.
In the intervention group, the systematic nursing intervention was implemented for the patients. The intervention group started to use a CPM machine and applied programed cryotherapy intermittently within one hour while returning to the ward on the day of surgery. The CPM machine was set to move from 0 degrees of extension to 60 degrees of flexion. The application of programed cryotherapy was continued for 20 minutes and then stopped for 30 minutes. The programed cryotherapy with the cryotherapy pack was replaced every four hours. From POD 1 to the day of discharge, the application of CPM and cryotherapy were conducted as the ward routine practice. The participants were blinded after assignment.
For the control group, a routine nursing procedure was conducted, but the CPM device and cryotherapy were not applied on the day of surgery. The patients started CPM from 0 degrees of extension to 60 degrees of flexion and received cryotherapy on POD 1. However, the application of cryotherapy was not programed, which meant the frequency and intervals of cryotherapy were not specified and determined by the patients or caregivers.
The patients, outcome investigator, and statistician were blinded. Because of the different frequency of the replacement of the programed cryotherapy on POD 1, it was impossible to blind the surgeon and the floor nursing staff. The following participant characteristics (age, sex, body mass index, religion and education), surgical variables (surgical experience, anesthesia method, wound length), the drainage amount, and the use of patient-controlled analgesia were collected.
Primary and secondary outcomes
A numeric rating scale (NRS) with 0–10 points  was the primary outcome of interest in this study. It was evaluated when 0 is no pain and 10 the worst imaginable pain. The secondary outcome included the short-form McGill Pain Questionnaire (SF-MPQ), ROM and the swelling status of the extremity. The SF-MPQ has been used to describe pain, feeling, memory, and influence in 15 pain situations . The SF-MPQ has been validated to have a high correlation coefficient with long-form MPQ. A universal goniometer was used to measure ROM of the knee joints. The measurement of the thigh circumference was performed 15 cm proximal to the superior pole of the patella with a measurement tape, meanwhile 15 cm distal to the inferior pole of the patella for the calf circumference. Both thigh and calf circumference were compared to the normal contralateral leg to determine the amount of postoperative swelling. The above measurement was conducted at POD 1 and POD 4 or until the patients were discharged. All the patients completed the outcome evaluation by a research assistant.
The sample size and Cohen’s effect size were calculated with the G*Power software package (version 3.1.4). The effect size f was interpreted as: f = 0.10 small effect, f = 0.25 medium effect, f = 0.40 large effect. The collected data were coded with numbers and input to computer software. According to the research purpose and variable characteristics, IBM SPSS Statistics for Windows (version 22.0; IBM Corp., Armonk, NY, USA) was used for information analysis. The statistical methods included descriptive statistics and deductive statistics. Descriptive statistics involved: (1) frequency and percentage, including category and ordinary variables of basic features of the investigated subjects, such as gender, education background, religious belief, and past surgical experience; and (2) mean value and standard deviation, including the equal-interval variables of basic features such as sex, education background, religious belief, operation experience, and anesthesia. Also, the scale score was described before and after measures. Deductive statistics included the independent t-test and Chi-square test before measures, as well as the homogeneity test of basic features in the intervention group and control group. An independent t-test was conducted to compare the pretest difference between the two groups, and a t-test of the samples was conducted to compare changes in the pretest and post-intervention groups. ANCOVA (analysis of covariance) was used to compare pain after the treatment, as well as differences in the posttest scores of knee flexion and knee swelling. A p-value < 0.05 was considered significantly different.