Effects of Respiratory Muscle Training in Mechanically Ventilated Adults

Patients who have been under mechanical ventilation for 48 hours are invited for participation in this study. Once participants are identified and his/her legal representative has decided to participate through an informed consent, subjects are randomly assigned to the experimental or conventional group.

Data collection begins with registering individual information into a data collection form that has been designed for the study; muscle strength is measured as Maximum Inspiratory Pressure (MIP). Then, patients are assigned either to experimental or group control using opaque sealed envelopes. Stratified randomization is performed, using SEPSIS status as strata, and within each category randomization block size is 4 and 6 respectively. Once sepsis status has been established, green envelopes are opened for patients with sepsis, and blue envelopes are opened for patients without sepsis. Finally intervention is performed according to group assignment.

Interventions

Conventional Therapy Group Control group receives conventional intervention of Cardiopulmonary Physical Therapy which has been established by institutional protocols, and consists of Cardiopulmonary Physical Therapy every 6 hours, physical therapy and mechanical ventilation management

Measuring Respiratory Muscle Strength

Respiratory muscle strength is measured by Maximum inspiratory pressure (MIP) following American Thoracic Society (ATS) and European Respiratory Society (ERS). protocols. MIP measures are performed in both groups, experimental and control. To measure MIP, we ask the patient to do a maximum inspiration and then hold it for 5 seconds; this inspiration begins from functional residual capacity (FRC). We will perform 3 measures with each patient and register the highest value obtained, using a Manuvacuometer for this measure.

Experimental Group

The experimental group will receive conventional treatment plus a respiratory muscle training protocol. Respiratory muscle training utilizes a Threshold device (IMT from Respironics Inc;Murrysville, USA), this device has been validated both in healthy subjects and in chronic obstructive pulmonary disease (COPD) patients

Study Setting

This study is being conducted at the Adult Intensive Care Unit in Valle del Lili Foundation, a private health institution level IV, located in the south of Cali, this is a referral centre for the southwest of Colombia.

Data Collection

Instrument design and pilot study

The researchers designed data collection tool to register data from clinical records and intervention, this tool is being used for both groups. The data collection tool includes demographic, clinical and intervention variables.

A pilot trial in a group of patients with similar characteristics to the study sample was conducted and then some aspects of the data collection tool and measuring protocols were adjusted

Research team and field personnel training

Four physical therapy specialists who had more than 4 years of experience in intensive care were recruited as monitors and data collectors.

In order to standardise and to determine validity of MIP measuring protocol, principal researcher conducted a training session with the physical therapists and then intra and inter-rater validity in 30 healthy subjects was established.

The primary source of information was the data collection tool, which was filed during intervention with each subject. The secondary source of information is the electronic medical record, recorded data are reliable because they are a basis for decision making in critically ill patients.

Patient recruiting and monitoring is performed by four physical therapists mentioned before. Data collection occurs from Monday to Sunday between 8:00 a.m. to 6:00 pm. Main researcher is not involved in data collection or intervention.

The variables obtained on patient admission to the study are: type of intervention, age, gender, medical history, diagnosis at admission, Severity Index (Apache II), hours of mechanical ventilation, weaning hours, cumulative dose of sedative received, cumulative dose of analgesics received, cumulative dose of relaxant received, cumulative dose of steroids, socioeconomic status and educational level.

The variable obtained during follow-up is: respiratory muscle strength.

The variables obtained at the end of follow-up are: Number of sessions of respiratory muscle training received, mechanical ventilation time (in hours), total hours of weaning, cumulative dose of sedatives, cumulative dose of pain medication received, cumulative dose of steroids, received physical therapy, number of physical therapy sessions, pulmonary rehabilitation received, number of pulmonary rehabilitation sessions, Non Invasive Ventilation requirement, pulmonary rehabilitation number of sessions, death, reintubation requirement, reintubation date, number of extubated hours because of reintubation, tracheotomy, tracheotomy date, septic shock, date of septic shock, death, date of death, decision not to continue in the study (date).

Data Quality

A monthly checking of data collection tools is performed by the principal investigator, using a checklist and comparing data recorded in the tool with medical records of subjects, this procedure is done before transcribing information on to the study database. Additionally, every data collection tool is reviewed using a checklist that includes three quality indicators: days between follow-up completion and closing of tool, number of empty boxes, number of diagnoses that are inconsistent; this procedure is also a responsibility for the principal investigator. Monthly meetings with the study team are held to talk about recording faults and how to improve quality of data collection and registration.

Database design and management

Once the data in the data collection tools have been checked, information is registered in an Epi Info database and then this will be exported to STATA 12 for analysis. Data registering in the Epi Info database will be done after every data collection tool has been closed and checked by the main researcher. Data registering will be performed by a person hired and trained for these procedures. Database management and analysis is performed by the principal researcher.

Data Analysis Plan

Patient recruitment and following-up are described using a flowchart designed according to CONSORT statement 2010. A preliminary exploratory data analysis will be conducted to evaluate data performance and basic assumptions required to implement statistical tests (normality, linearity and homoscedasticity). This analysis will also allow us to identify errors and biases in data collected, as well as outliers and missing data.

Sociodemographic and clinical characteristics of patients at study entry will be analysed descriptively according to measurement level of every variable using central tendency, dispersion measures, frequency tables and 95% confidence intervals. An analysis of clinical and sociodemographic characteristics of eligible patients who met inclusion criteria but were not included in the study will also be performed, as well as the reasons for not being included; this analysis will allow us to identify selection bias.

Finally univariate and multivariate analysis will be performed according to study objectives. Analysis will be performed on an intention to treat basis for all patients.

Respiratory Muscle Strength (MIP): Presuming that there is data normality, at univariate analysis, a difference between final MIP means of experimental group and control group will be analysed using a t-student. Differences in MIP means within every group will be analysed using linear regression. If normality, error independency and homoscedasticity are not met, we will consider data transformation for MIP variables. Once linear regression is performed a diagnostic test will be done to evaluate if the model represents or mimic observed data. In order to control for potential confusing factor an adjustment of variables will be done, we are considering as confusing factors: i)Cumulative medicament doses received before intervention, ii)Cumulative medicament doses received, iii) Time in ventilation mode before intervention, iv)Time in every ventilatory mode during intervention, v) Admission diagnoses. Additionally we will evaluate effect modification between variables mentioned before.

Ventilation time and weaning time: we will perform univariate analysis of total weaning time and total ventilation time between experimental and control group using a survival analysis (Kaplan-Meyer). We also intend to do multivariate analysis of total weaning time and total ventilation time using Cox regression adjusting potential confusing factors that we considered with other variables. We will verify Cox model assumptions about hazard proportionality during following-up. Need for reintubation: We are doing univariate analysis of need for reintubation using a comparison of proportions with Chi-square or Fischer exact test. Need for reintubation Relative Risk (RR) will also be determined for patients in both groups (Control and Experimental). We are doing multivariate analysis for the need for reintubation with logistic regression adjusting for potential confusing factors mentioned before.