Inspiratory Strength and Respiratory Complications After SCI Injury (RESCOM)

Background:

A spinal cord injury (SCI) often leads to paralysis of respiratory muscles (1), which causes a decrease in inspiratory volume and consequently limits the pre-cough volumes needed to produce an effective cough (2). The ability to cough is essential for airway clearance (3). With reduced cough capacity, the risk for pneumonia is increased (1). Pneumonia is still among the leading causes of death in the SCI population (4, 5). In acute SCI, up to 80% of patients are affected by respiratory complications (6) and in 51 % of these cases pneumonia occurs (7), corresponding to a cumulative pneumonia incidence rate of about 40%. Unfortunately, predictors of pneumonia in SCI are still poorly understood (6).

To the investigators knowledge, there are three studies on predictors of pneumonia in the SCI population (8, 9).

Two of these studies concentrated on factors of injury and shock severity (9) or identified the level and completeness of injury as the fundamental clinical entity predicting pulmonary complications (8).

However, both studies did not investigate potentially modifiable factors. The third study focused on potentially modifiable factors, such as lung function but did not cover respiratory muscle strength and did not define the inclusion of respiratory complications precisely (10). Therefore, to facilitate future evidence-based strategies to reduce the pneumonia risk in acute SCI, prevention targeted studies that aim to identify modifiable risk factors for pneumonia are urgently needed.

The current research proposal is targeted at specific parameters of respiratory function that are conceivably determinants of pneumonia risk as well as modifiable by respiratory muscle training (11,12). In a recent study (EKNZ 2014-293; accepted for Publication in Respiratory Care) the investigators evaluated various respiratory function parameters and identified inspiratory muscle strength as a reliable discriminator between individuals with and those without pneumonia.

Although the investigators identified inspiratory muscle strength as best parameter of respiratory function to discover individuals with SCI at risk for pneumonia, these data are based on a retrospective singlecenter data analysis without any detailed information on the type, time-point and severity of pneumonia. Also data on mortality due to pneumonia and other personal parameters that may influence the risk for pneumonia as well (e.g. smoking, exercise, mechanical ventilation dependency etc.) are missing in this analysis.

Objective:

The main objective of this study is to evaluate 'cut-off levels' of inspiratory muscle strength as predictor for pneumonia in individuals with SCI (diagnostic accuracy). Additionally, to evaluate lifestyle and treatment parameters as well as personal- and lesion characteristics as further potential determinants of pneumonia as well as the impact of pneumonia on quality of life and mortality.

Determination of sample size:

Regarding the in- and exclusion criteria, the investigators made a calculation of a realistic number of subjects who can be recruited for this study per year and center, based on the number of patients admitted to the corresponding center during the last years. Overall, the investigators will be able to recruit about 250 patients per year for this study.

With an anticipated number of 625 subjects over a 2.5 years recruitment period, the investigators may detect about 250 cases of pneumonia (under the assumption of a 40% incidence rate (6,7). Moreover, taking the diagnostic accuracy (i.e., area under the curve (AUC of 0.86 from a retrospective study as reference, the envisioned sample size of at least 250 cases per diagnostic group should allow for detecting a diagnostically and clinically relevant effect (i.e., change in AUC) of 0.07 to 0.05 with 95% confidence level and 80% power (18).

Criteria for withdrawal:

Reasons for which a participant needs to discontinue from the project:

• withdrawal of informed consent
• non-compliance
• medical complications which do not allow measurements of this study
• death

In case of drop-out, the already recorded data will be used for further analyses. All data of drop-out patients will be anonymized after analysis.

The investigators will report all reasons for withdrawal of the participants and compare the reasons qualitatively.

The study will not be stopped in case of futility, unless CTU Nottwil during the course of safety monitoring advices otherwise. In this case, CTU Nottwil will discuss potential stopping for futility with the study investigators group.

The final decision to terminate the study is taken by the Principal Investigator of the study RESCOM.

Risk-Benefit:

Patients have to perform a maximum of 4 measurement sessions (dependent on the time of rehabilitation) with 15 min of measurements each. Therefore the time needed for this study will not negatively influence rehabilitation of included patients. Respiratory muscle strength measurements and cough capacity measurements are non-invasive and also part of clinical routine. Safety of these measurements is very high and to the best of our knowledge the investigators are not aware of any adverse Events due to such measurements.

Potential benefit of project result:

Increased knowledge on determinants and effects of pneumonia will help adapt rehabilitative strategies/therapy to prevent patients from pneumonia and thereby also increase quality of life of the patients and life expectancy.

Methods of minimising bias:

Throughout the study, the coding of the participants will be conducted by the study nurses of each site in order to keep the data management and the biostatistician blind against the study condition as long as the data base is open.

The coding list remains with the study nurses of each site for the whole duration of the study. Thus, coding will be conducted without any influence of the principal investigator or raters.

The study investigators will strive for complete separation of the individuals involved in the steps before enrolment from those involved in the data management and analysis.

Assessments of primary and secondary outcomes will be conducted by a study nurse without any influence on data analysis. All study nurses will go through a profound assessment training program.

A study nurse will enter data into the secuTrial® data management system so that the biostatistician can analyze the data without having access to information about the origin, or names of participants.

Assessment of safety and reporting:

An annual safety report will be submitted once a year to the local Ethics Committee via local Investigator.

For multi-center studies the annual safety report contains information from all sites including information from sites outside of Switzerland. The Sponsor-Investigator prepares it, and then submits it to the participating Investigators. The participating Investigators submit it to the local committees.

Statistical considerations:

To evaluate the diagnostic accuracy of markers of lung function (particularly MIP) for predicting or excluding clinical pneumonia the investigators will use contemporary receiver operating characteristic (ROC) analyses techniques that allow for adjusting of covariate effects. Univariable and multivariable regression modelling will be used to evaluate the primary outcome and secondary outcomes. Multivariable modelling as compared univariable modelling will principally allow the investigators to investigate relative contribution of different predictors (or risk factors) to the outcome of interest, while controlling for potential confounding by other factors. Multivariable models may further be used to examine effect modification (interaction) as well as mediation.

Missing data:

Regression analyses of the primary outcome and secondary outcomes may suffer from missing data, which may cause bias and loss of information in analyses that are restricted to study participants with complete data ("complete-case analyses"). To appropriately account for missing data in advanced regression analyses the investigators will use multiple imputation techniques as well as an iterative imputation method (missForest) based on a random forest. Due to the relatively high anticipated sample size, the analysis will be split up into specific sub-Groups (e.g. completeness of lesion, para-/tetraplegia, type of pneumonia). For mortality flexible parametric modeling of survival data) and other secondary outcome analyses will be performed.

Data handling and record keeping/ archiving The secuTrìal data management system (iAS, Berlin, Germany), a web-based data capture and management system, will be used. The system is hosted on a server of the Swiss Paraplegic Centre.

The project set-up and eCRFs will be tested by the Data Manager of the Swiss Paraplegic Center, Clinical Trial Unit prior to the release into the productive environment using a custom testing protocol.

Every staff member is authorized for any CRF entry and by initials noted on the CRF the investigator can be subsequently identified. The data on the CRF will later be entered in the secuTrial database (iAS, Berlin, Germany).

One person will enter the data from the CRF paper form into the secuTrial database. A second person will control the correctness of the data in the secuTrial and the principal investigator will finally release the data for archival storage.

The Clinical Trial Unit will only give access to 2 people per center: Person A will get access for data entering and controlling and the local investigator (Person B) for data release.

Plausibility rules will be established in the data capture system to promote data quality. Furthermore, data will be reviewed and verified prior to data entry completion.

Encrypted data will be exported from the data management system for analysis in statistic software.

Data protection:

Project data will be handled with uttermost discretion and only be accessible to authorized personnel.

Direct access to source documents will be permitted for purposes of monitoring, audits or inspections.

The data management system allows to define roles for data entry, verification, validation and management. An audit trail documents all entries and changes made in the system and by whom.

Personal data of participating subjects which allow to identify an individual, will not be stored on a server or personal computer, nor will any names of subjects be transmitted to the main study center in Nottwil. These data will be printed on paper and stored in a locked data storage room with limited access in each participating center. An anonymous identification number will allow to connect the data entered in the system, but only by each center for their own subjects. Only anonymous data will be stored in the electronic data capture and management system. Encrypted data will be exported from the data management system for analysis in statistic software.

For data verification purposes, authorized representatives of a competent authority (e.g. an ethics committee) may require direct access to parts of the medical records relevant to the project, including participants' medical history.