Conventional Versus Automated Bag-based Mechanical Ventilator to Support ARDS Patients

The translation of novel technologies for Intensive Care Units (ICU) including mechanical ventilators has been mostly about incremental improvements through increasing features assuming the presence of highly trained personnel and dedicated infrastructure such as permanent power supply and high oxygen flow. Despite the long history of use of mechanical ventilators, the Covid-19 pandemic demanded innovation to meet worldwide demands for this type of life support equipment. The inadequacy of the commercially available ventilators has been mostly attributed to weak fabrication and supply chains and inconsistencies between the required sanitary infrastructure and human resources available for which most commercial ventilators have been designed. Diverse approaches addressed the design and development of low-cost and more available mechanical ventilators worldwide achieving various levels of clinical implementation.

In Peru, the Masi (companion in Quechua) ventilator was developed based on concepts of the MIT Emergency Ventilator (MIT E-Vent) by automating a resuscitation bag. The Masi initiative showed promising pre-clinical results that allowed the production of more than 250 units under an exceptional permit to be used clinically in Peru only when no other ventilators were available. The key features of this technology are its low manufacturing cost, low oxygen consumption, flexibility between non-invasive and invasive ventilation, and extended range of working environmental altitudes compared to commercially available ventilators. Clinical characteristics, blood chemistry information, respiratory parameters, and survival rates were analyzed for 77 patients (42 ventilated by Masi, 35 ventilated by commercial ventilators).

This is an observational retrospective cohort study. The study protocol was approved by the Research Ethics Committee for Life Sciences and Technology of the PUCP (002-2021-CEICVyTech/PUCP) on September 23, 2021. A total of one hundred and three (103) patients were admitted to the ICU between January 19, 2021 until August 27, 2021, which coincides with the second Peruvian wave of the Covid-19 pandemic in the Hospital Vitarte II in Lima, Peru. Data from patients who were diagnosed with Covid-19, with signs of Acute Respiratory Distress Syndrome (ARDS), intubated, and connected to a mechanical ventilator in the ICU were de-identified for this study. From the 132 variables collected into the Electronic Health Records, only 120 selected variables were transcribed into an electronic database (Microsoft Excel V). The data was standardized and entered by a physician trained in Good Clinical Practice (Data Entry Investigator) following a Data Dictionary.

After data cleanup, this study retrospectively analyzed data from eighty-seven (87) patients who had a positive Covid-19 diagnosis and used a mechanical ventilator machine. According to the Peruvian regulatory agency, DIGEMID, Masi was indicated only when no conventional ventilator was available, and a patient needed one as determined by the ICU physician. The physicians used their criteria to prescribe Masi or other available ventilator brands.

During the described period, the ICU of the Hospital Vitarte II had seven beds and four conventional ventilators: Evolution (EVL100001-T-NF), Vela (16532-07), Newport (e360) and Dragger (Savina 300). Eight new individually validated ventilators known as Masi were provided to the ICU, and the medical personnel were trained regarding its clinical use. This device was approved to be used in Peru during the sanitary emergency by the local competent authority: Dirección General de Medicamentos, Insumos y Dispositivos Médicos (DIGEMID, Spanish acronym) based on preclinical evidence of safety.

To avoid bias derived from the convenience sampling of a single-site study, the Data Entry Investigator reviewed the medical charts corresponding to every patient (n=103) admitted into ICU at the site during the study range across Masi before selecting the charts to be included. A selection was made from the initial 132 variables recorded based on relevance for the present study. Variables from the initial database were uploaded into R studio. The variables and values were reviewed by two independent data analysts and queried to the data entry assistant when necessary. The dataset was cleaned up, and variables were normalized and categorized. Each variable was evaluated with a univariate analysis followed by a bivariate analysis comparing different results between pre- and post-intubation, and post and last-intubation measurements. The main factors used for comparison were the respiratory parameters before and after the intubation between the mechanical ventilators Masi versus conventional. An additional analysis was done to compare survival rates by using the Kaplan-Meier estimator and COX analysis. A sensitivity analysis was done to each model to evaluate suitability for a non- parametric and semi-parametric test.