The design of the study was developed in accordance with the Consolidated Standards of Reporting Trials (CONSORT statement).
This study will take place at multiple sites (i.e. Zurich/CH (leading center), Barcelona/E, Toledo/E, Heidelberg/D, Murnau/D, Nijmegen/NL and Glasgow/GBR). Local Ethics Committees at each center have approved (Barcelona, Toledo and Zurich) the study. Informed consent will be obtained from all subjects prior to participation. Patients will be included as they are referred to one of the participating centers (consecutive sample).
Patients with a subacute traumatic SCI initially categorized as ASIA B or C with a motor level between C4 and T12 and who are only partially able to walk (Walking Index for Spinal Cord Injury-WISCI ≤5 ) will be eligible. Subjects should be able to start the training within 60 days after trauma. Patients who do not comply to the requirements of the Lokomat training device (i.e. bodyweight >130 kg, body height >200 cm, leg length diff >2 cm, osteoporosis, instable fracture of lower extremity, restricted range of motion, decubitus ulcer of lower extremity) or with concomitant injury limiting walking ability (e.g. lower extremity fractures, instable spine fractures, Joint instability preventing weight-bearing, severe soft tissue lesion, traumatic brain injury) or with pre-existing medical conditions interfering with unrestricted walking (e.g. total joint replacement, chronic pain, osteoarthritis, polyneuropathy, cardiopulmonary disease) or who are older than 60 years or younger than 18 years will be excluded from participation. Patients who already participate in other rehabilitation or pharmacological study will also not be considered for participation.
In previous studies with stroke subjects [18–23] the mean difference in walking speed amounted to 0.0418 m/s. This value was used to calculate the sample size further assuming a standard deviation of 0.05 m/s, a statistical power of 0.8 and a significance level of 0.05. The calculation resulted in the requirement of 23 subjects in each group to be able to reject the null hypothesis.
Patients will be randomly assigned to either the intervention or the control group using a computer generated 4-block randomization scheme. The allocation will be performed by an independent person not otherwise involved in the study. The responsible researchers at each center request the group allocation by mail.
The locomotor training with the Lokomat device should start within approximately 30 days but not later than 60 days after the SCI. The observation period of the training for this study lasts 8 weeks. For the initial 5 trainings there are no defined specifications. These trainings serve to optimize the setup and for the patients to familiarize with the robotic locomotor training device. During subsequent trainings following guideline will be adopted to adjust the training to the actual capacity: body weight unloading will be reduced to the least tolerated amount (no knee buckling or toe dragging). The walking speed will be set within the range of 1.6 to 3.1 km/h and the guidance force in the range from 100% (full assistance) to minimum tolerated. The training session will be shaped in the following way: 3 min walking without specification (warm-up period) after that, every 3rd minute either speed, visual feedback or guidance force will be changed. This will avoid that the training becomes monotone and lacks challenge.
Patients who will be assigned to the intervention group receive one or two Lokomat trainings per day on 3-5 days per week. The Lokomat walking time per day should not be shorter than 50 min.
Patients of the control group receive one Lokomat training per day on 3-5 days per week. The Lokomat walking time per day should not be longer than 25 min.
Outcome data will be assessed in the respective centers by therapists and medical doctors. Hereafter the data will be sent anonymously to the PI for analyses. Due to the characteristic of this study neither the therapists nor the assessors nor the patients can reliably be blinded regarding the group allocation. The principal investigator (PI) who is the first author of this report (MW) and will be involved in the analyses will not be aware of the respective intervention.
In order to describe the characteristics of the included subjects, demographic and clinical data will be assessed. The primary outcome is the self selected walking speed using the Ten Meter Walking Test (TMWT) . It will be assessed at baseline, bi-weekly during training, and at the six months follow-up. In addition the following items will be assessed at baseline, bi-weekly during training, and after six months: the Walking index for spinal cord injury -WISCI  (an ordinal scaled index for the assessment of walking capabilities with 21 categories. Zero represents that the patient is not able to stand or walk, the maximum of 20 means that the patient can walk without bracing, walking aids or personal assistance), the maximum walking speed, the ASIA classification , the detailed Spinal Cord Independence Measure-SCIM , the modified Ashworth scale-MAS  of hip and knee joints, and the Penn spasm frequency scale . During the training period the mechanical stiffness and the maximum voluntary torque of the legs as well as the cooperation of the patient during the training will be assessed bi-weekly using the force transducers of the robotic training device. For every training the distance walked, walking speed, walking time and body weight unloading, the rate of perceived exertion as well as the occurrence of any events (e.g. skin breakdown or joint stress or scheduling problems) will be assessed. At the end of the training period, i.e. after 8 weeks the patients subjective impression about the success of the training will be assessed using the Patients' Global Impression of Change Scale-PGIC .
Since at the time of the study all patients will undergo a rehabilitation program, they receive along with the Lokomat training the usual rehabilitative therapy (i.e. physio and occupational therapy). In order to assure that the standard rehabilitation program is comparable between the participating centers, therapy schedules of one week will be collected of four patients per center. These schedules should contain the amount of therapy and roughly the content of the therapies (e.g. focus on arm or leg or standing and walking or other activities).
All analyses will be performed in the leading center by the PI. He receives the data in form of an electronically completed case record form after completion of the training. For the analyses, the data will then be incorporated into the statistical software (PASW statistics 17.0, IBM-SPSS Inc. Chicago/IL).
In order to describe the characteristics of the sample, descriptive statistics will be applied. For the analysis of training induced differences between intervention and control group, T-Test for independent groups or the non-parametric correspondent depending on the outcome variable will be applied. In addition, multilevel models will be used to evaluate potential confounding factors (e.g. center effects, age differences, neurological level of lesion or ASIA) as required.