The SCIentinel study - prospective multicenter study to define the spinal cord injury-induced immune depression syndrome (SCI-IDS) - study protocol and interim feasibility data

The “SCIentinel” trial is designed as a prospective multicenter study for the detailed evaluation of the SCI-IDS (Figure 1). The overall coordination is performed by the Department of Experimental Neurology, Clinical and Experimental Spinal Cord Injury Research (Neuroparaplegiology) at the Charité University Hospital, Campus Mitte, Berlin, Germany.

The study comprises seven recruiting trial centers with a specialization in SCI treatment: Treatment Center for Spinal Cord Injuries, Trauma Hospital Berlin, Germany; Center for Musculoskeletal Surgery (Campus Virchow Clinic); Department of Trauma and Reconstructive Surgery (Campus Benjamin Franklin); Department of Neurosurgery (Campus Virchow Clinic and Campus Benjamin Franklin), Charité University Hospital Berlin, Germany; Division of Trauma Surgery, University Hospital Zurich; Spinal Cord Injury Center, University Hospital Balgrist, Zurich, Switzerland, and the Department of Neurosurgery, Toronto Western Hospital, Toronto, Canada. The study investigators are physicians specialized and experienced in the treatment and rehabilitation of patients with SCI.

A recruitment period of 30 months is scheduled. Each patient will be followed-up to three months post-trauma. After completion of recruitment and follow-up, a further six months period is planned for the database clearing, the statistical evaluation and the preparation of the trial report. Enrolment has started in August 2011. Expected study completion date is the first quartile of 2014. Publication of the trial report is scheduled for the end of the year 2014.

The protocol received approval by the local Ethics Committees: Ethical Committee of the Charite – Universitätsmedizin Berlin (EA1/001/09), University Health Network Research Ethics Board, Toronto (REB10-0384-AE), Cantonal Ethics Commission, Zurich (KEK-ZH-Nr. 2011–0059). Participants will be informed about the trial, orally and in written form, using patient’s information sheets and written informed consent will be obtained. This study complies with the Helsinki Declaration, the principles of Good Clinical Practice (GCP) and the Personal Data Protection Act. The study will also be carried out in keeping with local legal and regulatory requirements. The study has been registered in the German Clinical Trials Registry (DRKS-ID: DRKS00000122).

A number of 118 patients is planned to be allocated to the study for examination of the primary endpoint. In order to allow for investigation of the role of disturbed innervation of the major splanchnic nerve (T5-T9) as linked to alterations of the immune system function, we categorize SCI patients into lesions with the neurological level T4 or above versus lesions neurological level T5 or below. Therefore, the patients will be assigned to 3 groups: i) 48 SCI-patients with a neurological level T4 or above; ii) 31 patients with a neurological level T5 or below; iii) 39 patients with an acute vertebral fracture, but without SCI as control group (Figure 2).

The study will not interfere with methylprednisolone treatment per site regimens and patients will be also enrolled, but are excluded from the analysis of the primary endpoint. Based on experiences from the clinical practice in the recruiting trial centers the rate of cases treated with methylprednisolone is estimated to be 5-10%. Thus, up to 12 patients will additionally be included for explorative examination of methylprednisolone-related alterations of the immune system after SCI (Figure 2).

The primary endpoint is a difference in HLA-DR expression on monocytes exactly quantified by analysing antibody-binding molecules/cell within the first week (day 3–4) post-trauma. Numbers of patients to be enrolled were calculated on the basis of a pilot study (n = 26) [6]. The sample size calculation was performed for an effect size of 0.18. The type one error was be set to 0.05 (two-sided); type two error was set to 0.2. The number of patients eligible for evaluation (for testing the hypothesis by the one way anova) is 56 in total (software nquery, version 7.0). Due to an expected rate of drop-outs/missing data of 50% and non-parametric testing, where applicable, the number of patients to be recruited increases to 118 in total. The expected rate of drop-outs/missing data relates to the structural conditions of acute SCI care in terms of i) delayed transfer to specialized SCI trial centers after primary care in external non-specialized centers and ii) loss to follow up due to discharge from inpatient rehabilitation, particularly in the control group or in cases of mild incomplete SCI. Consequently, to allow for a completion of the study within a feasible time period consideration of missing values is unavoidable.

The sample size calculation was performed taking the asymmetric natural distribution of the neurological lesion level into account [20]. Injury to the cervical and high thoracic spinal cord has the highest incidence and represents the largest group of SCI patients. Thus, the relation of the 3 groups to each other was determined as follows based on the pilot trial data i) 41% (n = 48); ii) 26% (n = 31); iii) 33% (n = 39).

With the seven study centers treating 20 to 60 acute SCI patients admitted to the emergency ward of each center per year we estimate to screen about 200 SCI patients per year and about 40 of them are expected to meet eligibility criteria (Table 1).

Investigators will evaluate the patients after admission to the emergency ward for eligibility. Before final inclusion into the study the investigator will conduct an interview with the patients to verify the inclusion- and exclusion-criteria and obtain written informed consent.

Measurements and observations are scheduled in all groups in the same manner (Figure 3). An inclusion assessment will be performed at the study entry and is documented in a case report form (CRF). The baseline CRF includes the neurological classification, assessment of injury date and time, medical history, concomitant injury, acute SCI therapy concerning high-dose methylprednisolone treatment as well as surgical intervention.

At the 5 study visits, equal blood samples and clinical data will be obtained. The clinical data comprise infections, medication, surgery, and blood transfusions. The first assessment should be completed as early as feasible after the injury, preferably before surgery, otherwise at least within 31 hours post-injury. The second assessment point refers to a time window of 31–55 hours post-injury. To ensure a sufficient time interval to the first assessment the second one should be performed on the following day at the earliest. This allows a better intra-individual analysis. The visits 3, 4 and 5 refer to day 7, 14 and to 10 weeks after injury, respectively. Blood withdrawals, except at time point 1 should be performed between 7:00 and 11:00 am. This setting determines the influence of the circadian rhythm to a minimum.

In the case, if it is unavoidable that a patient may be assigned to the study later as 31 hours (delayed recruitment) e.g., due to later admission or transfer to a study center (see also sample size calculation), study inclusion is tolerated at any later time point. In individual cases of delayed recruitment or patient’s absence, documentation of visits 2–5 may be scheduled in extended time windows (Figure 3).

The most prevalent infections after SCI, in detail chest infections and urinary tract infections, are diagnosed and documented in the study according to established definitions of disease to ensure a biometric comparability [15, 21, 22] (Table 2). Additionally, a distinction is made between non-symptomatic and symptomatic urinary tract infections, i.e. the latter are classified with regard to evidence of fever, increased spasticity, bladder spasm, suprapubic or flank pain, autonomic hyperreflexia, malaise, or lethargy. All other infections are documented according to the usual diagnostic criteria of the participating centers.

The neurological evaluation is performed according to the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI), a revision of ASIA classification [23, 24]. Within the ISNCSCI-regime the ASIA impairment scale (AIS) as a measure for completeness of the injury, and the single neurological level of the lesion are assessed.

Peripheral blood is collected under sterile conditions from each participant at each visit. All samples are labeled with a six-digit pseudonym and any personal information of the participants is removed. Overall, one 2 ml and one 6 ml Ethylenediaminetetraacetic acid (EDTA) BD Vacutainer®, two 3 ml Heparin BD Vacutainer® and one 8.5 ml Serum BD Vacutainer® are needed. Furthermore one 5 ml Cyto-Chex® blood collection tube (BCT) and two 2.5 ml PAXgene™ RNA tubes are collected (Figure 4). Blood samples for flow cytometric analysis are analyzed as soon as possible, preferably within 24 hours latest within 48 hours. Preservation of intracellular and extracellular surface markers is ensured by using Cyto-Chex® BCT [25, 26].

Supernatants from whole blood stimulation and further blood samples for secondary outcome measures are stored at -80°C for subsequent batch analysis (Figure 4). In the trial centers at Berlin and Zurich, one additional 8 ml BD Vacutainer® CPT™/Ficoll™ tube is collected for immediate preparation of Peripheral Blood Mononuclear Cells (PBMCs) for storage in liquid nitrogen. The handling of samples follows identical standard operating procedures and checklists at all trial centers.

The following investigations are performed to determine the immune status: i) immunophenotyping with lymphocyte subpopulations (B, T, NK cells), T cell subpopulation (CD3, CD4, CD8) and monocytic HLA-DR expression; ii) ex vivo T lymphocyte cytokine secretion such as ConA-induced IFN-g, TNF-α, IL-2, IL-4, IL-5, IL-10 secretion and staphylococcal enterotoxin B (SEB)-induced IL-17 production; iii) gene expression analysis for immune activation/deactivation related genes (mRNA expression); iv) circulating immune modulators: such as IL-6, epinephrine; v) indirect infection parameters: procalcitonin, C-reactive protein (CRP), LBP (lipopolysaccharide-binding protein).

Immunophenotyping using fluorescence-activated cell sorting (FACS) including quantitative measurement of monocytic HLA-DR expression as the primary outcome measure will be conducted according to standardized operation procedures established at the Institute of Medical Immunology (Charité - Universitätsmedizin Berlin, Germany) for clinical diagnostics of an immune suppression state [14]. HLA-DR belongs to the MHC class II molecules that are responsible for antigen presentation to CD4+ T-cells. Monocytic HLA-DR expression reflects the state of the adoptive immune competence as it is positively regulated by Th1 cytokines and negatively regulated by anti-inflammatory cytokines, stress hormones and some exogenous agents. Furthermore, monocytic HLA-DR expression is associated with the ability to produce pro-inflammatory mediators upon challenge by bacterial products such as lipopolysaccharides (LPS). Reduced monocytic HLA-DR expression has been established to correlate with an increased risk for infections in critically ill patients previously [14, 27, 28].

In contrast to the pilot study that measured HLA-DR expression on different cell types [6], the SCIentinel study measures quantitatively the amount of HLA-DR molecules per monocyte (in estimated numbers per cell). The reliability of the method has been previously approved in an inter-laboratory study [14].

The principle of the test is as follows. A mixture of Phycoerythrin (PE) beads conjugated with defined amounts of PE molecules is measured at the same instrument settings as the cells incubated with a mix consisting of anti-human HLA-DR-PE, anti-human CD14-PerCP-Cy5.5, and an inhibitor of HLA-DR turnover (BD Quantibrite™ HLA-DR/Monocyte reagent, BD Biosciences). The PE beads facilitate conversion of the FL2 axis into PE molecules bound per cell. The known ratio of PE to anti-HLA-DR antibody is used to convert the PE molecules per cell into antibodies per cell (AB/c). The anti-HLA-DR antibody, clone L243, reacts with a non-polymorphic HLA-DR epitope and is conjugated with PE molecules in a ratio of 1:1. The anti-CD14 antibody, clone MοP9, is conjugated with PerCP-Cy5.5. CD14 is expressed by the majority of monocytes. In addition, since the cyan dye component of PerCP-Cy5.5 binds to CD64, the anti-CD14 PerCP-Cy5.5 antibody detects all monocytes (CD14 brightly positive and weakly positive). The immunodiagnostical thresholds delineating a graded SCI-IDS penetrance are defined as indicated in Table 3 and were developed based on previous clinical trials after CNS injury [27].

Data are collected on a paper CRF (pCRF) basis. All patient’s data are managed with a six-digit pseudonym. At the coordinating center personnel familiar with the trial protocol check the filled pCRFs for completeness and consistency. Implausible or missing data may be corrected or added after consulting the investigator at the trial site through the study management (Queries). The corrected documents will be archived together with the completed CRFs. Data are entered and stored electronically in a database (Access) and are independently double-checked for correctness. The database has secured and restricted access centralized at IT-Center of Charité-Universitätsmedizin Berlin, Germany. Data backups are performed on a daily basis.

An administrative interim evaluation for control of the estimated recruitment schedule after enrolment and follow up of approximately 50% of the calculated sample size has been performed. Patients recruited until April 2013 and followed up to July 2013 were included. The groups for comparison were analyzed for their concordance with the estimated sample size (Table 4). Furthermore the completeness of data for the main laboratory outcome parameters (flow cytometry) at each time point was assessed within the 'per protocol’ population (Table 5).

Data analysis of the primary endpoint consists of a comparison, between the 3 groups at the early and subsequent follow-up assessment points using one-way ANOVA or nonparametric tests, if appropriate. The analysis will be performed using the full dataset comprising all patients included according to the criteria as defined in the study protocol. In addition, in patients with complete intra-individual datasets also changes over time within the groups will be assessed with descriptive methods. Furthermore, in the case of a relevant amount of missing values, multiple imputation will be taken into consideration for verification of the analysis of the original data. Data obtained from patients treated with methylprednisolone will be primarily analyzed in a casuistic manner.

The statistical analysis of the administrative interim evaluation has been performed using SPSS for Macintosh, Version 19.0.

In order to differentiate the SCI-related 'neurogenic’ effect on the immune system from the 'non-neurogenic’ effect of the surgical intervention/trauma itself (activation of the stress axis/post-aggression syndrome) the study protocol incorporates a control group of patients with isolated vertebral fracture without neurological deficit. Of note, this refinement of the control group definition by the SCIentinel trial overcomes a weakness of the pilot study where patients after smaller surgical interventions such as arthroscopy served as control group [6], which only insufficiently mimics the post-aggression syndrome caused by surgical stabilization of the spine.

In order to evaluate the impact of the lesion height we assess differences between rostral and caudal spinal cord lesions. Consequently, SCI patients were divided into two groups with/without expected impairment of the major sympathetic outflow and compared with the controls applying a three-group strategy for analysis (Figure 1). To assess the effect of lesion severity, we compare complete (AIS A) with incomplete lesions (AIS B-D) in a descriptive sub-group analysis.

We defined the eligibility criteria (Table 1) in order to minimize inclusion bias and to obtain 'representative’ data. The definitions stipulated in the SCIentinel protocol are consistent with previously published studies [6, 8, 18, 30]. It was necessary to exclude patients with severe, life threatening polytrauma or severe TBI in order to limit variability caused by concomitant injury. For detection of immune system functionality, it is furthermore essential to exclude all circumstances combined with pre-existing alterations in the immunological profile such as chronic autoimmune diseases and pre-existing systemic treatment with corticosteroids.

The administrative interim analysis of the SCIentinel trial has confirmed the feasibility of the eligibility criteria. From August 2011 to April 2013 n = 64 patients have been enrolled for the study which represent 37 patients per year. Thus, the enrolment status is within the expected range. Four of the 64 patients (6%) were treated with high-dose methylprednisolone. Therefore, those patients are to be excluded from the 'per protocol’ analysis. However, casuistic studies on immunosuppressive effects of methylprednisolone treatment in the context of the SCI-IDS are enabled. The remaining 60 patients represent 51% of the estimated sample size (Table 4).

Observational time points (Figure 2) range from acute to sub-acute stages after SCI comparable to the pilot study [6]. Here, the time schedule corresponded to the hallmarks of parenchymal inflammation following SCI. The implementation of sub-acute assessments is in line with previous trials that demonstrated mid- and long-term alterations of the immune system after SCI [8, 18, 19].

The setting of defined time corridors instead of time points provides a more complete tracking of the participants. In order to prevent a high amount of missing datasets due to loss to follow-up, the protocol allows an inclusion at each documentation time point.

The administrative interim analysis of data completeness for the main laboratory parameters indicates - depending on the group or visit – a completeness of data between 24% and 91% with critical rates of completeness at visit 1 and visit 5, particularly in the control group (Table 5). As a consequence, the recruiting centers have been requested to enroll the patients as soon as possible after injury. Additionally, we intend to address the challenge of loss to follow up by travelling to external rehabilitation centers or by phone calls to remind patients of their study visit at the outpatient clinic. The rate of complete longitudinal datasets ranges between 18% and 36% but is expected to increase during the further trial course since trial routines are better established. Noteworthy, the sample size calculation takes 50% of missing values into account and was performed for one-way ANOVA. Repeated measure analysis requires lower sample sizes. Therefore, it seems feasible that a relevant amount of complete longitudinal datasets can be achieved also for the description of the intra-individual course of the SCI-IDS. Pursuing this study concept, serial investigations on short- and mid-term changes of the immunological function are provided.

For the clinical research, it is imperative to use reliable reproducible clinical tests. The ISNCSCI standards revised over the years to provide better, more specific definitions [23, 24]. We incorporated the AIS and the single neurological level into the trial protocol for neurological baseline characterization of SCI patients.

The determination of infections is of crucial importance in this study. Prevailing infectious complications of SCI patients are pulmonary and urinary tract infections [4]. Cameron and colleagues recently provided a review consisting of 12 articles of urinary tract infection (UTI) screening for SCI [31]. Noteworthy, there is no universal definition of UTI used in the literature allowing substantial heterogeneity. The American Paraplegia Society (APS) recommends to base the diagnosis of UTI solely on bacteriuria, with a threshold of 10² colony-forming units (cfu)/mL for SCI patients [32]. However, up to 75% of samples taken from asymptomatic SCI male patients under intermittent catheterization contain ≥ 10² cfu/mL [32]. Thus, a threshold of 10² cfu/mL is associated with low specificity, leading to an excess of treatment, with the risk of the emergence of resistant bacteria, adverse events, and unnecessary expense. In consequence, we defined according to the National group on urologic rehabilitation of paraplegics urinary tract infection as bacteriuria greater than 10⁵ cells/ml or WBC count of more than 100/mm³[21].

Regarding the development of chest infection criteria defined by Mann et al. were incorporated [22]. For the diagnosis of pneumonia the finding of infiltrates or opacity in chest X-ray will be required consistent with Haeusler et al. who used similar criteria in the analysis of cellular immunodepression preceding infectious complications after acute ischemic stroke [15].

In summary, the clinical assessments and definitions used in the SCIentinel study allow for collection of reliable and reproducible data.

For meeting the purpose of a more comprehensive characterization of the SCI-IDS and establishing prognostic surrogate parameters, it is essential to implement specific parameters, readily validated in clinical immunology, robust and reproducible [14]. Quantification of immune cell populations in the blood using FACS is a clinically established and precise method that qualifies as primary outcome measure. Our primary measure is the determination of HLA-DR-molecules per monocyte. Alterations in HLA-DR expression on monocytes occur early enough to identify patients at high risk for relevant immune depression and consecutive development of infections. Significantly decreased monocytic HLA-DR levels have been observed as early as day one after acute cerebral ischemia in patients who developed infectious complications in comparison to patients with an uncomplicated clinical course [27]. Furthermore, the method is already established and available in centralized immunological laboratories. It has sufficient diagnostic specificity as a basis for clinical decision-making and therefore is clinically highly relevant. Moreover, we implemented a large setting of secondary endpoints for the development of very early predictive parameters, for example on the level of mRNA expression.

The objective of this study is the longitudinal characterization of immunological markers of the innate and acquired immune system after traumatic SCI. Furthermore, the aim is to detect diagnostic surrogate parameters for the clinical application after SCI. The characterization of a causative relationship between the lesion level and/or SCI severity and the type and/or extent of the immune response after SCI - including a critical characterization of the methylprednisolone effect - is essential for the implementation of therapeutic studies. Understanding and recognition of the immunological dysfunction and the altered susceptibility to infections may then assist in consecutive decision-making. This is relevant for the stratification of patients who are at high or low risk for infectious complications [33]. In addition, these findings may also hold the key for early therapeutic immunomodulation aiming to improve overall survival by anticipation and prevention of life-threatening infectious complications.