Effectiveness of specific stabilization exercise compared with traditional trunk exercise in women with non-specific low back pain: a pilot randomized controlled trial

Ethics

The Clinical Ethics Committee of University Hospital Sant Joan of Reus approved this study (12-06-28/6proj4). All participants gave written informed consent before data collection began.

Study design

A pilot randomized trial was conducted in Catalonia from February 2013 to February 2015 (NCT02103036). Participants, diagnosed by medical practitioners and referred for treatment of non-specific LBP, were randomized using computer-generated random number tables into two treatment groups: a TTEP group and a SSEP group (Fig. 1). Afterwards, measurements were taken at baseline (session 0), at half intervention (session 10), post-intervention (session 20) and one month later. A single-blind study was conducted, due to the impossibility of achieving double blindness—the physiotherapist performing the intervention had to know which treatment each participant was to receive. The therapist who performed the intervention was a qualified health professional with 5 years’ experience in the field; a different professional took all the study measurements, he was blinded to the participant’s assignment. Participants were blinded to their group allocation, design and hypotheses (Page & Persch, 2013).

Participants

Participants entering the trial were required to meet the following inclusion criteria: females aged between 18 and 70 years; diagnosed with non-specific LBP (fewer than 6 weeks of pain duration) by a specialist doctor who used imaging such as magnetic resonance, radiographic or computed axial tomography to rule out other spinal disorders, and under no pharmacological treatment for pain. Exclusion criteria were: diagnosed with other spinal disorders and/or any other serious co-morbidities (e.g., cancer, severe lung pathology); presence of cognitive impairment; inability to perform exercises; having followed a specific training program with a physiotherapist in the previous three months; having been treated with analgesic infiltration in the previous 6 weeks, or failure to follow their 20-treatment schedule exactly. All study participants were volunteers, and all underwent intervention under the Faculty of Medicine and Health Sciences of University Rovira i Virgili.

Intervention

Both treatment groups (TTEP and SSEP) underwent 20 sessions of treatment at a frequency of three to five sessions per week (Saner et al., 2011), as follows:

In the first five sessions, the only treatments were application of an infrared lamp and transcutaneous electrical nerve stimulation (TENS), since these have been proven to reduce pain in both acute and chronic LBP (Bertalanffy et al., 2005; Jauregui et al., 2016; Van Tulder et al., 2000). Patients did not perform any exercise in these first sessions, since there is strong evidence that exercise is not effective in relieving acute pain, and can even worsen symptoms (Van Tulder et al., 2000). TENS therapy was applied with a multichannel portable TENS unit (Megasonic 313 P4, Carin) on the lumbar spine. Biphasic square wave impulses at a frequency of 100 Hz and pulse duration of 70 µs were used for a total duration of 20 min. Four rectangular 90 × 45-mm electrodes were applied on the fascia thoracolumbaliis and approximately 10 cm proximal to this, along the midline of the muscle (i.e., directly over the site of pain) (Kofotolis, Vlachopoulos & Kellis, 2008).

In sessions six through 20, each group engaged in its respective back school exercise program in regular sessions (Fig. 2); they followed 30-minute protocols of 10 exercises, with 10 repetitions of each. The TTEP group performed exercises from the LBP protocol developed by the Physiotherapy and Rehabilitation Service at Sant Joan University Hospital (Reus, Spain). The SSEP group performed exercises gathered in a search of literature on core stability exercises (Koumantakis, Watson & Jacqueline, 2005; Shamsi, Sarrafzadeh & Jamshidi, 2015). Before beginning exercises, participants received education on the anatomy of the back (members of the SSEP received a simplified explanation of core musculature), correct posture, and spinal alignment, as part of a back school.

The same physiotherapist supervised all classes. Upon treatment completion, he gave each participant a home exercise program which included the exercises from their treatment sessions.

Outcome measurements

The primary outcome was pain, measured with a 10-cm Visual Analogue Scale (VAS), which has been shown valid and reliable. VAS is a numerical rating scale (0 = no pain to 10 = worst imaginable pain) which represents the intensity of the current pain and allows the evaluator to compare it with previous or later evaluations (Hawker et al., 2011). VAS was used to measure pain at baseline session 0 and at sessions 10 and 20. Pain measurement one month after the final treatment session was done by telephone, using the 11-point Numerical Rating Scale (NRS). During this phone conversation, each patient also performed a verbal, subjective assessment. Both these final assessments are considered sufficiently sensitive to detect clinically relevant pain changes (Hawker et al., 2011); they are even considered interchangeable for calculating pain in lumbar pathologies (Hawker et al., 2011; Thong et al., 2018; van Tubergen et al., 2002).

One secondary outcome was disability, measured using the Roland Morris Disability Questionnaire (RMDQ). The questionnaire asks 24 questions related to the participant’s current functional status. Different studies have shown RMDQ to be a useful and reliable instrument for evaluating participants with LBP (Payares, Lugo & Restrepo, 2015). RMDQ was used to measure pain at baseline session 0, at sessions 10 and 20, and one month after the final treatment session.

Another secondary outcome was degree of inflammation, as measured by blood-sample levels of the cytokines interleukin 6 (IL-6) and tumor necrosis factor alpha (TNF-α). These markers were used because they have been found to play significant roles in relation to back pain (De Queiroz et al., 2016; Kraychete et al., 2010). The presence of some inflammatory mediators might be associated with pain and disability in patients with LBP, since pro-inflammatory cytokines such as IL-6 or TNF-α contribute to the activation of nociceptors that generate potential of action and pain hyper sensibility (Cui et al., 2000; Queiroz et al., 2015). Degree of inflammation was measured at the beginning and end of the treatment program by enzyme-linked immunosorbent assay (ELISA) method. Blood samples were collected by a qualified doctor (blinded to group allocation) and will be obtained from the antecubital vein (Tomazoni et al., 2019).

The study recorded additional factors, including anthropometric characteristics (age, height, weight, body mass index [BMI]), and degree of physical activity using Quick Classifier of Physical Activity (ClassAF) in Metabolic Equivalent of Tasks (METs). ClassAF is a global questionnaire which classifies people as physically active or inactive using a corresponding qualitative formula (Vallbona et al., 2007). All these data were collected before the intervention began by the trained physiotherapist.

Data analysis

Groups were compared with respect to change, from baseline (session 0) to half-intervention (session 10), baseline to post-intervention (session 20), and baseline to 1 month after the intervention concluded; from session 10 to session 20 and session 10 to one month post intervention; and finally from session 20 to one month post intervention.

SPSS program version 23 Windows was used to analyze the data. A descriptive analysis was made of the study sample, with standard averages, deviations and percentages of the different variables collected. The Kolmogórov-Smirnov test was applied to assess data distribution in each group. A Student-t test was done to assess differences between the two treatments, and effect size was calculated to measure the magnitude of the experimenter effect, using the standardized mean difference (SMD) for variables normally distributed and the effect size of Mann–Whitney’s U test for variables not normally distributed (Field, 2005). Two-way repeated ANOVA analyses were used to examine differences over time. Assessments were carried out using non-parametric tests for variables that did not present normal distributions. The level of statistical significance for the study was established at p < 0.05.

Flow of participants and therapists through the trial

59 potential participants were referred to the research team. Of those referred, 20 were not included, for various reasons (Fig. 1). 20 participants were placed in the traditional trunk exercise group (TTEP); the remaining 19 were placed the specific stabilization exercise group (SSEP). Of these 39 participants, 30 completed their course of treatment (15 from each group). Table 1 shows participants’ baseline characteristics: age, height, weight, BMI, physical activity level, pain or disability. According to the CONSORT statement, significance testing of baseline differences in randomized controlled trials were not performed (Moher et al., 2010).Two members of the SSEP group could not be reached for the one-month follow-up telephone call.

Compliance with the trial method

30 (76,9%) participants attended all 20 intervention sessions. Once the intervention was completed, the physiotherapist advised participants to repeat their exercises at home, three times a week for one month. 15 (50%) participants reported performing the exercises as advised; 9 (30%) reported doing their exercises occasionally; 4 (13.3%) did not perform exercises at home; the remaining 2 (6,67%) were unreachable.

Effect of intervention

Data on pain and disability are shown in Table 2; data on degree of inflammation are in Figs. 3 and 4.

Results show an insignificant effect size and no significant differences between groups in terms of current pain intensity, or for any outcome measure. At the end of intervention (session 20), pain intensity for the TTEP group had decreased by 0.33 cm (95% CI [−1.7–1.0], p = 0.615) more than in the SSEP group. Both back school treatments showed positive results for pain reduction from baseline to end of treatment, and baseline to one month post-intervention. In the TTEP group, pain (baseline to final session) reduced by 4.6 cm (95% CI [3.3–5.8]); the SSEP group’s reduction was 4.3 cm (95% CI [3.0–5.6]).

Similarly, there were an insignificant effect size and no significant differences between groups in terms of change in disability. At post-intervention (session 20), disability levels in the SSEP group had decreased by 0.40 points (95% CI [−1.7–2.5], p = 0.701) more than the TTEP group. Both back school treatments yielded positive results in disability reduction, baseline to end of treatment, and baseline to one month post-treatment. In the TTEP group, RMDQ scores reduced by 5.1 points (95% CI [3.0–7.3]) from baseline to post-intervention. In the SSEP group, RMDQ reduction for the same interval was 6.1 points (95% CI [3.7–8.6]).

Figures 3 and 4 show outcomes for inflammation. TNF-α showed higher values for the TTEP group than the SSEP group at the two visits where TNF-α was measured. Significant differences were observed between the groups, baseline and post-treatment. In the first case, a difference of 66.97 pg/mL (95% CI [6.3–139.5]) was recorded; in the second case the difference was 128.94 pg/mL (95% CI [52.8–205.0]).

In contrast, IL-6 levels were found to be similar between the two treatment groups, with no significant differences observed. At baseline the difference was 2.51 pg/mL (95% CI [−2.3–7.3]); at post-intervention the difference was 1.25 pg/mL (95% CI [−3.9–6.4]).

In reference to the evolution of inflammatory biomarkers between baseline and post-treatment, the results for participants who practiced traditional TTEP indicate an increase in TNF-α levels of 46.16 pg/mL (95% CI [13.0–85.3]) and a tendency toward decreased levels of IL-6, 0.19 pg/mL (95% CI [−1.6–1.2]).

In contrast, the results in the group that practiced SSEP are the other way around: there was an increase in IL-6 levels of 1.06 pg/mL (95% CI [0.03–2.1]) and a tendency toward decrease in TNF-α levels of 12.81 pg/mL (95% CI [−42.3–16.7]).