Medicine in spine exercise (MiSpEx) for nonspecific low back pain patients: study protocol for a multicentre, single-blind randomized controlled trial

A pilot feasibility study was conducted (unpublished material). The therein found outcome with the largest effect size was selected as the primary outcome for the present study. Thus, primary outcome is maximal back extension torque. Sample size calculation was based on the effect size of d = 0.24 (based on the changes in primary outcome from baseline to post-intervention (12 weeks)). For the main study, effect size was conservatively completed to d = 0.2.

Participants are allocated to intervention or control group at the ratio of 2:1. Normal distribution and variance equality are assumed. Following a determined α-error probability of 2.5 % and a β-error probability of 0.1, a minimal sample size of 1186 patients are consequently to be included into analysis (n = 791 in the intervention and n = 395 in the control group). As a dropout rate of 30 % is assumed, a total of N = 1542 participants are included in the study. Although motor control and stabilizing exercise are likely to show equal effects for the general population and for top athletes [15], medical access and the potential to implement medical exercise regimes into everyday life differ between these populations (e.g. training schedule, competitions). According to study aim 1, and beside the studies’ main target group, we make additional effort to also recruit a sample of athletes to gain further (beyond the main analyses) insight and test the feasibility in this much less investigated population. For dichotomization purpose, only top competitive athletes (A to D pool, participating in top level international, federal or state competitions) are further treated as athletes.

Representative volunteers are recruited at clinical low back pain consultation hours, through flyers, local newspapers and bulletins, and personal recruitment. Athletes are further recruited by personal approach in one of the participating study sites, mostly at the mandatory annual sports medical examination. Only general information is provided during recruitment. Interested persons are scheduled for a visit in one of the six study sites and then screened for eligibility. Males and females aged 18–65 years with chronic low back pain are considered eligible if they fulfil the following inclusion criteria: at least one episode (≥ 4 days) of nonspecific back pain in the last 12 months, ability to understand the extent and meaning of the study and to answer a questionnaire without help.

Exclusion criteria consist of acute infection, pregnancy, inability to execute a one-legged stance, diseases with contraindication for exercise (including spine complaints based on space-occupying, inflammatory, traumatic or systemic processes [12]) and change or progression in the severity, localization or type of back pain during the last 7 days. Following application of inclusion and exclusion criteria, patients are approved by the physician in charge. Each participant signs informed consent prior to study enrolment.

After inclusion and completing the first visit, volunteers are block-randomised (n_block = 18; 2:1 basis) into either the intervention or control group by the study coordinator. Randomization order follows the study inclusion order. Each study site receives its own randomization lists once a week from the primary study centre; the randomization sequence is generated using a computer-based algorithm (www.​randomization.​com). All assessors and study personnel other than the study coordinator (and the therapist) are blinded. Participants are told not to communicate their group allocation to other participants or study staff.

At each measurement day (M1-M5) a psychometric assessment is performed to detect the moderating factors (‘yellow flags’) for the chronification of back pain. The test battery will be rotated (Fig. 1) and is partially based on subscales in order to reduce the burden for participants. Pain is assessed by the Chronic Pain Grade questionnaire (CPG) [16], which allows a differentiation in the subscales pain intensity (PI: 0 = “no pain” to 100 = “the worst pain imaginable”), and disability (DS: 0 = “no disability” to 100 = “I was incapable of doing anything”) in the past 3 months. Furthermore, participants can get classified into one of the four hierarchical pain and disability grades ranging from low pain/disability (grade I) to high pain/disability scores (grade IV) [17].

The five moderating factors were assessed using the following psychometric instruments: (1) physical activity: International Physical Activity Questionnaire (IPAQ) [18] and type of physical activity. (2) Pain experience: pain-related cognitions by the Fear-Avoidance Beliefs questionnaire (FABQ-D, [19]), avoidance‐endurance‐behaviour by the Avoidance-Endurance Questionnaire (AEQ; [20] and anxiety and depression by the Hospital Anxiety and Depression Score (HADS-D; [21]) ). (3) Stress: chronic stress by the Trier Inventory of Chronic Stress (TICS; [22]) and critical life events in the last 3 months. (4) Personal life context: social support by the Berlin Social Support Scales (BSSS; [23], individual attachment style by the Relationship Questionnaire 2 (RQ‐2; [24, 25]) as well as the socioeconomic status (CASMIN; [26] and lifestyle (alcohol, smoking, general health and sleeping status). (5) Medical care context: health services options (see also DSF, [27]) .

Postural control in single and bipedal stance is assessed using balance boards (Wii Balance Board, Nintendo, Kyoto, Japan; modified by CSMi Computer Sports Medicine Inc., Stoughton, MA, USA). Postural sway is estimated by the trace length [mm] of the excursion of the centre of pressure (COP). For the measurement, participants were asked to maintain in upright stance as still as possible for 30 seconds without wearing shoes, with their hands on their hips and their eyes open, standing on two balance boards. Afterwards, the single leg stands (left and right) followed in the randomized order (each 30 s, eyes open, one single balance board) with the contralateral knee bend (90°).

In comparison to a laboratory-grade force platform and with an intraclass correlation (ICC) of 0.81 (single limb stance, 95 % CI 0.39–0.92) and 0.77 (double limb stance, 95 % CI 0.69–0.95), respectively, the Wii balance board shows a sufficient validity [29]. Accordingly and with an ICC of 0.89 (95 % CI: 0.76–0.95) for single limb balance assessments and an ICC of 0.86 (95 % CI: 0.54–0.9) for bipedal stance, a high reproducibility was shown [29].

Jumping ability is afterwards determined using the same balance boards. Countermovement jumps (CMJ) and combined CMJs followed by reactive jumps are executed. Whilst each participant performs double leg jumps, single leg jumps are, for security purpose, performed by athletes only.

After warm-up (30 s tapping on a stepper with a 14–21 cm high) and familiarization, participants perform two bipedal CMJs, followed by two combined bipedal CMJs/reactive jump. For the reactive jumps, all participants are told to immediately add another jump after landing from the CMJ. All athletes afterwards jump twice with the left and twice with the right leg (randomised order), each time starting with a CMJ followed by a CMJ with a reactive jump. Main outcomes will be jumping height in cm for the CMJ and contact time [ms] for the reactive jumps.

The balance board is able to precisely (r = 0.99 in comparison to gold standard laboratory-grade force platform) measure vertical ground reaction force and thus jumping performance [30]. Data from both the posturography and the jumping tests are stored for the later analysis using the HUMAC Balance program (CSMi Computer Sports Medicine Inc., Stoughton, MA, USA).

A non-invasive external three-dimensional wireless movement analysis system (IMU sensors, Ergotest Innovations, Porsgrunn, Norway) with a 200 Hz sample rate is used to collect kinematic data. Sufficient test–retest reliability of this measurement system for lumbar spine kinematics (r² > 0.97; RMSE = 4.1°; p < 0.001) has recently been demonstrated [31]. Six sensors are attached, two on the lower back (each 5 cm left/right from L3), two on the pelvis (left and right iliac spines), and two on the thigh, each 10 cm above the medial tibia plateau. Participants perform over the maximal range of motion in a self-determined and comfortable velocity with eyes open. After familiarization trial and device zeroing, all subjects perform two times maximal spine flexion-extension movements, followed by maximal lateral flexion movements (two times), followed by two maximal rotation movements (side order randomised). Maximal ranges of motion (ROMs) [°] will be selected for further analysis.

Data were stored on a data processor for the following offline analysis. Maximal ROM was computed from raw data using MuscleLab Professional (Ergotest Innovations, Porsgrunn, Norway).

The Timed Up and Go test (TuG) for mobility testing and the Chair Rise Test (CRT) for lower extremity strength/force are performed during clinical function testing [32]. For the TuG, subjects will sit on a chair with arm support, stand up, walk straight 3 m to a mark on the floor, turn around, walk back to the chair and sit down again. Participants then (CRT) are set on a chair (46 cm sitting height, arms crossed in front of the breast) and are asked to stand-to-sit five times as fast as possible without losing control.

Time [s] is the outcome further analysed for CRT and TuG; common stopwatches are used to assess it. Both tests are performed twice; the best result will be selected for further analysis.

Maximum trunk extension strength assessment with isokinetic or, depending on the participating centres’ isokinetic device availability, isometric maximum trunk extension torque testing [Nm] is considered to be the primary outcome. To obtain maximal torque (isokinetic, n = 4 study sites) or maximum isometric voluntary force (isometric, maximal isometric voluntary force (MIVF), n = 2 study sites) of the trunk extensors (primary outcome) and flexors, validated devices from various but nevertheless renowned producers are used.

After a specific warm-up with 30 isokinetic repetitions at 60°/s (isometric: two submaximal practice trials), five (three) tests at 60°/s (3 s) were performed, separated by 2-min rest intervals. Ranges of motions (isokinetic) are device-specific set to 40–45° in flexion and 10–15° in extension. Subjects were verbally encouraged in a standardized way to elicit maximal effort. Collected data were immediately analysed and the highest torque (force) was considered to be representative for further statistical analysis.

All interventions are guided (centre-based) or instructed (home-based) by trained and experienced medical training or sports therapists. Following the randomization at the end of M1, intervention group participants are scheduled to a 3-week centre-based guided intervention, followed by a 9-week home-based individual training. In each phase and in accordance with the guidelines [33], intervention participants are told to train three times a week with a mean duration of 30 minutes. The program consists of four different sensorimotor exercises. For individualization purposes, all exercises comprise 12 different levels and offer the possibility to add self-initiated additional motor tasks like ball tapping (single-handed, on the floor or against the wall and/or additional weight). Two of the exercises are dedicated to directly train the core stabilizing and/or core surrounding muscles whilst the two other exercises are considered to impact indirectly on the upper and/or lower extremities. The exercises are commonly described as: (1) quadrupedal/all-fours stability; (2) deadlift/rowing; (3) double leg–single leg heel-pad stance; (4) side planks. Exercises will be performed with three series of ten repetitions each. In between series, a 2-minute break (self-controlled) is held. Details on the intervention and belonging levels are provided in Table 1.

For individualization purposes, the level (including the number and type of self-initiated additional motor tasks and additional weight) for each of the four exercises is determined by an experienced therapist at the beginning of the intervention for all participants in the intervention group. The therapist in charge rates the participants’ performance accuracy at level one and derives a starting level. Performance accuracy is thereby standardized rated based on the axis and plane alignment (extremities, trunk) during motion, movement goal (endpoint) accuracy and no loss of balance during motion or single movements order. The starting level is defined as the highest level in which this accuracy can be reached. This level (including the number and type of self-initiated additional motor tasks and additional weight) is further adaptive and may be corrected in both directions (increase or even decrease in level) continuously by the therapist during the centre-based phase. The goal for the intervention is to increment by one level once a week until the maximum level 12 is reached.

The intervention group patients receive individualized exercise counselling for home-based training as follows: at the end of the centre-based trainings, the starting level for the home-based phase is determined (therapist in collaboration with the participant). In addition, the goal level for the home-based phase is determined. Each subjects’ goal level and the periodization to reach this level is subsequently recorded on the training log. Intervention group subjects were additionally encouraged to contact the counselling exercise therapist by phone, e-mail or in person at any time during the home-based intervention. Final therapy individualization may be accomplished by using a standardized procedure to detect potential non-responder and/or non-complier. This detection is based on cutoff values which will be determined from the studies’ questionnaires. If a potential non-responder and/or non-complier is detected, each study site is recommended to implement an additional behavioural module to the intervention. This module and its empirical funding is described in detail elsewhere [34].

Control group participants do not receive any additional intervention in/from one of the study centres but may proceed with their regular care (e.g. physiotherapy or manual therapy) or additional interventions (the same applies to the intervention group) if they are engaged in one at the time of study inclusion. The intervention within medicine in spine exercise (MiSpEx) thus can be seen as additive.

During the stationary centre-based phase, a standardized training log is completed by the therapist, documenting the exercise level and the (if applied) additional weight. To monitor home training compliance, intervention group patients are asked to fill in an exercise log during the home-based training (between M2 and M3).