A member of the research team generated a randomized list of 32 patients assigned to two groups. The randomization method used was by blocks. Free license software, recommended by Suresh25 and available at www.randomization.com was used. In this study, the number of patients per block was 4, with a total of 8 blocks, thus considering the 30 patients calculated for the sample size and two more that were added to balance the groups. 32 opaque envelopes containing a card with two possible legends were listed: A (experimental) or B (control). The envelopes were previously delivered to another member of the team (blinded), located in another hospital, who by telephone transmitted the assignment to a sports doctor (not blinded) who gave the participants instructions regarding the intervention that would be given to them (taping technique and exercise).

To determine the change in the degree of knee pain, the WOMAC scale was applied. This scale consists of three dimensions: pain, stiffness and functionality. It is graded with a Likert scale of 0–4 and a higher score denotes a worst situation (96 points – 20 for pain, 8 for stiffness and 68 for functionality). Several studies have been published about its reliability.26

The exercise intervention was based on the quadriceps strengthening program published by Chang et al.9 We developed a program in which the patients performed a dynamic-type strengthening exercise (6–8 per Omni Perceived Exertion Scale-Resistance Exercise Scale [OMNI-RES]) in their home, with a volume of 3 sets of 15 unilateral repetitions, with extension and flexion movement for both knees (2-s duration per movement), with a frequency of twice a day. For calculation of 1-Repetition Maximum (1RM), a red, green, or blue band test (Theraband®) was performed, requesting that the patient perform at least 10 repetitions of extension. The band was tied to a belt worn by the participant, which that would aid in stabilizing the exercise. The participant leaned against a wall, sitting on the floor with both knees in extension. They were asked to flex and adjust the elastic band to generate tension against the extension of the knee. The band was adjusted to the level of effort requested according to the OMNI-RES scale.13 With this elastic band, the participants performed, in their homes, 3 sets of 15 repetitions of extension and 3 sets of 15 repetitions of knee flexion unilaterally. Rest intervals between each series were 30s. Frequency of execution was 3 days per week. The patient was also asked to perform stretching exercises for quadriceps and hamstring muscles, lasting 15s per muscle group, twice a day and 6 days which lasted a week. In order to increase adherence to the intervention, a daily control register was issued for the participant's exclusive use, in which he/she would record the exercises they performed on the corresponding days, the time it took to do them, and whether there were any adverse events. The KT was installed according to the manufacturer's specifications (Kinesio®), contained in the envelope in which the pre-cut product is sold. This was done during each follow-up visit during the study. Installation was accomplished by a single applier, which is certified by the manufacturing company for a basic course and declares no conflict of interest. The kinesiotape pre-cut knee tape pack contains 3 sections of tape: 2 black (I-strip and Y-strip) and 1 blue (I-strip) tape. With the knee flexed at 90°, the base of the black “I” strip was adhered on the leg midline about 15cm above the interarticular line (IAL). No tension was applied and the strip was spread over the same midline to about 5 –7.5cm below the IAL. The Y-strip was then applied, also without tension, and each tail extended to the sides of the midline. After applying the KT to each participant, generation of the “convolution” effect on the skin was confirmed, requesting the participants’ complete extension of both knees. The placebo technique was installed under the same conditions, on days different from those of the participants of the experimental group. The same material was used, but without the specifications contained in the envelope: a single black “I” strip with high tension (>50%) and while the knees were flexed at 90° with a bearing surface. applications were carried out by means of an unblinded applier, given the nature of the intervention. To maintain the sample blind, in each evaluation participants were asked to wear pants or long skirts to prevent them from showing the applied technique. Participants were given an appointment after 1 week to reinstall the corresponding tape, check the daily control register, and confirm permanence in or withdrawal from the study.

In the first session, prior to application of the tape, the Visual Analog Scale (VAS) for pain intensity (0–10cm) and the WOMAC index were applied for baseline measurement. At the end of weeks 2, 4, and 6, the VAS and WOMAC were again applied. At the end of week 6, the notebook was reviewed and a final appointment was assigned to inform the participants of the results.

Our hypothesis was that the group with strengthening exercise and KT would have 2 fewer points of difference in knee pain, compared with the control group, measured through the WOMAC pain subscale. With a α of a 0.05 and a β of 80%, we consider data published by Chang et al.9 to express a mean difference of 1.7 points for pain (Standard Deviation [SD]=0.9 and 1.5) and of 8.2 points for physical function (SD=4.8 and 8.3) of WOMAC subscales, a sample of 12 participants was calculated per group. The sample was increased by 20% to prevent losses, and two more participants were added to balance the groups for the performance of block randomization. N was 16 women per group.

Descriptive analysis of demographic variables was performed. Baseline comparisons were made with the Student's t test and the χ2 test and the correlation coefficient for pain was carried out by the VAS and the WOMAC index. On intergroup analysis, a Friedman test was used for related samples, calculation of percentage change for EVA and WOMAC and recalculation of Spearman's correlation coefficient between pain variables during the comparison of repeated measures. Intergroup comparison was conducted as follows: calculating analysis of covariance (general linear model) with Bonferroni statistic; Hotelling trace criterion to determine the significance of interaction between times (repeated measures) per treatment group; percentage of change for VAS and WOMAC (from the baseline value of the covariable corresponding to the baseline mean of both groups at week 6 of follow-up), and correlation coefficient between VAS and WOMAC variables during comparison of repeated measures and by interaction between assignment type and BMI category on percentage of change in pain. Analysis was performed by intention-to-treat. The software used was IBM SPSS Statistics for Windows (Version 23.0; IBM Corp., Armonk, NY, USA) Significance level was set at p≤0.05 throughout the study.

This study took into account the principles of respect, beneficence, and justice, as well as international ethical guidelines with regard to recommendations for an additive design.27,28 The study was initiated once approval was obtained from the Institutional Review board and Ethics Committee. Informed consent was accepted and signed by each participant.

We interviewed 39 subjects who fulfilled inclusion criteria. 7 of them were excluded by diverse conditions. In the control group, there were 2 voluntary dropouts (change of address and labor issues) and in the experimental group there were 4 (two reactions associated with the use of the KT but not with the materials – use of an external fixative substance and adhesive tape–, voluntary dropout for academic reasons and other due to mechanopostural low back pain). We followed up the corresponding cases (Fig. 1).

Fig. 1.

Flow chart of the study participants.

(0.2MB).

In the baseline state, there were no significant differences in demographic variables, but the allocation groups differed in terms of mean stiffness (p=0.004), with a higher score in the placebo group (Table 1). In this baseline data, as might be expected, pain measured by VAS and WOMAC correlated with a Pearson r coefficient of 0.445 (p=0.01); BMI correlated with physical function (r=0.400; p=0.02). Within WOMAC, pain with stiffness (r=0.553; p=0.001) and pain with physical function (r=0.601; p<0.001) also exhibited correlation. Finally, stiffness and physical function correlated with a coefficient of r=0.363 (p=0.04). Only physical function tended to be different according to the degree of obesity, with 27.3±9.9 points for participants with grade I and 21.5±7.6 points for those with overweight (p=0.07). The remaining variables (age, VAS, pain, and stiffness) were similar between patients with obesity and those with overweight (p>0.05). There was also no association between degrees of obesity and the affected knee (p=0.28) or with degree of OA (p=0.32), and the latter were not associated with each other (p=0.53). Baseline VAS was higher in cases in which the affected knee was the right one with 7.4±0.9 points vs. 6.0±1.6 with the left one (p=0.008); pain and stiffness. On the other hand, Baseline VAS also differed according to the degree of OA: grade 3 was more painful (10.6±8.3 points) than grade 2 with 7.8±2.5 points (p=0.02), and there was greater stiffness, with 4.5±2.2 points for grade 3 vs. 2.9±1.3 points for grade 2 (p=0.02).

In the experimental group (Table 2), the VAS decreased from 6.4 to 4.6 (−28%) from baseline to week 6 (p<0.001), while pain decreased from 7.6 to 5.5 (−27.5%) during the same time period (p=0.01). There was no significant change in stiffness, and the physical function scale decreased −14.7% with respect to the baseline score (p=0.07). The percentage of change in pain correlated with percentages of change in stiffness (r=0.772; p<0.001) and physical function (r=0.693; p=0.003); therefore, as pain decreased, stiffness decreased and physical function improved. On the other hand, the change in pain and the decrease in stiffness correlated with a change in an increase in physical function (r=0.880; p=0.0001). In the placebo group (Table 2), VAS decreased from 6.7 to 5.2 (−22.3%) of the baseline value at week 6 (p=0.001), while pain decreased from 9.0 to 5.9 (−34.4%) during the same time period (p=0.0001); stiffness decreased from 4.0 to 2.9 (−25.7%) (p=0.003) and the functional scale was reduced by 26.0% (p=0.05). Similar to the correlations observed in the experimental group, in the control group, the percentage of change in pain correlated with the percentages of changes in stiffness (r=0.888; p<0.001) and in physical function (r=0.612; p=0.012), implying that, as pain decreased, stiffness decreased and functionality improved. In addition, the change in the decrease in stiffness correlated with a change in the increase of functionality (r=0.521; p=0.03).

The change percentage was calculated from the baseline value of the covariate, which corresponds to the baseline mean of both groups up to week 6 of follow-up. Respectively for placebo and experimental group, there was a difference of 8.2% improvement in the VAS (1.9 and 1.3 points, p=0.16), while for the subscales there were 1.1% for pain (2.7 points for both groups, p=0.2), 8.3% for stiffness (0.7 and 0.4 points, p=0.1) and 5% for physical function (4.5 and 5.7 points, p=0.65). None was significant. According to the Bonferroni test, no significant differences were observed in any of the measurements of weeks 2, 4 and 6, except for week 4 for the pain subscale (p=0.01) with a difference of 1.3 points and in week 2 for the stiffness subscale (p=0.04) with a difference of 0.9 points.