INTRODUCTION
The efficacy of non-invasive ventilation (NIV) in respiratory failure was described in the ‘90s. From that moment forward, it has been used to treat diseases that have traditionally been managed at the Intensive Care Unit (ICU), such as chronic obstructive pulmonary disease (COPD) and car diogenic pulmonary edema.1
The use of NIV can reduce the need for intu bation, shorten the hospital length of stay, and reduce mortality, resulting in a rational use of the resources.2,3 The acquired experience has allowed its use outside the ICU.4 The failure or success can depend on the type of patient, the staff’s skill, and follow-up tools. Each institution has to design its own response protocol according to its resources.4-6
Frequently, patients with chronic respiratory failure show exacerbations that require hospital ization.2-9 To decide where to deliver NIV inside the hospital can be a complex decision involving several factors such as the kind of underlying disease, associated comorbidities, the severity of the physiological damage, and the patient’s preferences.5-9
On the other hand, NIV is used in environ ments without continuous monitoring, such as the patient’s home (HNIV), where the ventila tory treatment is guided by the health care team and the family, for the purpose of improving the patient’s quality of life.5-10
Finally, a group of patients with progressive diseases or advance disease stage develop respi ratory failure. In these cases, NIV can be used in combination with other treatments to mitigate dyspnea.10,11
In 2017 we organized a multidisciplinary team to deliver NIV at the regular hospital ward in patients without immediate indication for ICU admission in the absence of clinical severity signs (according to the Plant12 criterion) or advance di rectives for therapeutic limitation.
OBJECTIVES
To describe the clinical profile of patients who received NIV in a regular hospital ward and their clinical outcome.
MATERIALS AND METHODS
Retrospective study in patients treated with NIV in a regular hospital ward.
The protocol was approved by the Ethics and Institu tional Review Committee of the Hospital Británico de Bue nos Aires, in accordance with the Declaration of Helsinki (protocol: CRI# 1052, March 2020).
Population
The study included consecutive adult patients who had been admitted to the regular ward of the Hospital Británico de Buenos Aires between January and December 2019 (12 months), and received NIV. They were included when they were receiving HNIV and were admitted due to an acute event, or if they were transferred from the ICU to the regular ward. Patients with tracheostomy younger than 18 years were excluded.
Clinical data
The following information was obtained from the unique electronic medical record (EMR), SAP™: medical history, reeason for hospitalization, and previous use of NIV. Spi rometries performed in our institution (MedGraphics Paul. Saint Paul, USA, Nhanes III reference equation) up to 3 months before were taken into account.
Our center is a clinico-surgical general hospital with 350 beds and 40 beds for adult intensive care. The Respiratory Kinesiology, Pulmonology, Internal Medicine and Intensive Care Departments have residents and personnel on duty 24 hours a day.
The indication and use of NIV was decided by a multidis ciplinary team. Patients were examined when they began the NIV and two hours later, with daily periodic visits in the morning and at night.
The ventilation parameters were selected after the process of adaptation, gradually, taking into account the clinical status until the patient achieved a balance between efficacy (objective) and tolerance (comfort and compliance).
Ventilators were classified in the following way:
- Basic, level I devices: continuous flow generators for barometric ventilation with basic compliance monitor ing, without battery or alarms.
- Intermediate, level II devices: continuous flow genera tors for barometric ventilation with battery and high priority alarms, with ventilatory efficiency and compli ance monitoring.
- Advanced devices with life support (level III): volume-controlled or pressure-controlled ventilation, alarms of different priority levels, availability of internal and external battery and full monitoring.
The ventilator choice was based on clinical complex ity (acute, chronic, palliative care), hours of use and vital support requirement. The interface was selected for each patient. The monitoring was obtained through a ventilator-integrated software; Encore Pro II ™ and Direct View ™ (Philips-Respironics™. Murrysville USA) and ResScan 10.1™ (ResMed™. San Diego, USA).
Follow-up included clinical examination and physiologi cal parameters (vital signs, ventilatory mechanics, state of consciousness, and pulse oximetry: SatO2). The basal arterial blood gas (ABG) was obtained in the morning. Hy percapnia was defined as PaCO2 > 45 mmHg. The current protocol proposes that gasometry should be performed daily in unstable patients during the first hospitalization stage, and every 48 h or whenever there is clinical change during the patient’s stay in the regular ward. We suggested daily data download from the internal memory of the ventilators in accordance with our institutional protocol, with review of efficiency and compliance and possible parameter ad justment. The respiratory polygraphy was made whenever there was inconsistent monitoring information, or in cases of lack of clinical improvement after multiple parameter adjustments.
We reviewed the clinical outcome (death, hospital dis charge, inclusion in a palliative care program, limitation of therapeutic effort), hospitalization place, use of ventilatory support, and referral to and from the ICU.
The decision to discharge patients from the hospital was made jointly between participant services, once the patient showed clinical stability, family and social support, organized home care (in case this modality was applicable) and gasometric improvement. Upon discharge, the patient coordinated daytime visits to hospital and outpatient offices of the corresponding specialist physician. Figure 1 shows the current follow-up protocol.
Statistical analysis
Categorical variables are expressed in percentages.
In the case of continuous variables with normal distri bution, results are expressed as mean and standard devia tion. Numerical variables without normal distribution are expressed as median and percentile (25%-75%).
Differences between groups were compared through the Mann-Whitney Test or the c2 Test for quantitative and qualitative variables, respectively. Comparisons including three or more groups were made through the Kruskal-Wallis Test and the non-parametric Cochran’s Q Test.
A p value > 0.05 was considered statistically significant.
The analysis was conducted with the Prism 8.02 software (Graph Pad, La Jolla, CA).
RESULTS
NIV was delivered to 43 patients of 68.3 ± 12.4 years with a body mass index (BMI) of 30.1 ± 12.3 kg/m2 for one year. 44.2% were obese and 67.4% showed hypercapnia. The male/female ratio was 1: 1.
The diagnoses were (n = %); COPD (17; 39.5%), neuromuscular diseases (16; 37.2%), obesity and hypoventilation (5; 11.6%), heart failure (3; 7.0%) and thoracic cavity restriction (2; 4.7%). Table 1 shows the characteristics of the population.
Reasons for admission were COPD exacerbation (39.5%), heart failure (21%), progression of muscle weakness (14%), pneumonia (11.6%), percutane ous gastrostomy (4.7%) and emergency surgery (4.7%), among other things (4.7%). Hospital length of stay was 12.1 ± 7 d for all the population; 14 ± 9 d for survivors; and 5.7 ± 3 d for deceased patients.
Twenty-eight patients (65.1%) were using HNIVat the time of admission. The remaining 15 patients (34.9%) began NIV at the ICU and were transferred to a regular ward after they were stabilized.
The barometric mode with backup frequency (S/T) was the most widely used (79%) (Table 2). Most patients (90%) used oronasal masks, and 21 (48.8%) needed supplemental O2, especially COPD patients (Table 2). The time NIV was use was 12.7 ± 10.2 d, with a compliance of 6.9 ± 3.1 h/d, and we could observe more extensive use in COPD (p < 0.01) (Table 2).
Twenty-nine patients had PaCO2 > 45 mmHg upon admission (67.4%); and 8 had pH < 7.35 (18.6%).
Three patients (7%) were admitted to the ICU due to clinical worsening (2 for impaired conscious ness and one due to progressive hypercapnia), though none of them died.
Thirty-five patients were discharged with HNIV (94.6% of dicharged subjects). There were 7 (20%) new indications of home ventilation (Table 4). The PaCO2 upon discharge was lower (46.1 ± 4.6; p 0.05), even though 10 patients (27% of survivors) were discharged with PaCO2 > 45 mmHg (Table 3).
Finally, 16 patients (37.2%) were included in palliative care programs (mortality of 37.5%: 6 patients); there were 5 COPD patients and 5 cases of metastasised prostate cancer. The deaths were COPD-related and had some aspects in common: patients were older and had advance directives for therapeutic limitation, high PaCO2 levels and took opioids to alleviate dyspnea or refractory symptoms (Table 5).
DISCUSSION
This analysis describes the use of NIV in the regular ward of a general hospital in Argentina, and shows that NIV was delivered to patients with COPD, neuromuscular disease, hypoventilation due to obesity and respiratory failure (especially hypercapnic). A significant proportion of patients was using HNIV or needed ventilatory support upon discharge.
Some hospitals have developed specialized re spiratory care units.13,14 Our hospital doesn’t have that type of unit, but we do treat a considerable amount of NIV candidates outside de ICU. Even though general ward rooms don’t have centralized monitoring, we established the following nurse/ patient ratio: 1:5, residents and kinesiologists with active care shifts, 24 h a day.
The number of ICU beds limits the number of patients who are admitted for respiratory failure. In their study, Lapichino et al showed that there is a tendency among intensive care physicians to give priority to surgical patients, mostly those with clinical diseases,15 particulary when unfa vorable results are expected. On the other hand, ICU admission of less severe cases or patients with chronic diseases implies risks (for example, infection, isolation, delirium, etc.) and increases healthcare costs.
NIV has been delivered at the regular hospital ward for more than one decade. In an Italian study, 56% of 756 patients were successfully treated with NIV (60% due to COPD exacerbation). Also, 47% of the subjects were directly referred from the emergency unit.16 In out center, patients who had already adapted to NIV had priority for admission to a regular ward. A small percentage of patients (7%) were transferred to the ICU.
In many parts of the world,14 the regular ward environment is considered inadequate for deliv ering NIV.4 However, a multicenter, randomized, controlled trial,18 directed by Plant et al showed the efficacy and safety of NIV in a regular ward in pa tients with COPD exacerbation and mild respira tory acidosis (pH 7.30-7.35). When NIV is delivered by qualified personnel, it reduces the number of occupied beds at the ICU, the intubation rate and mortality.18 In our series, due to preestablished safety criteria, the admission pH (7.36) was higher than the value reported in similar studies.14-18
COPD exacerbation has been the most analyzed disease for the use of NIV outside the ICU. It has been suggested that a diagnosis other than COPD would be predictive of failure. More experience is necessary so as to recommend NIV in a regular ward in other situations.
COPD patients showed high values of PaCO2 and hypoxemia and needed supplemental oxygen and higher-pressure support, though they didn’t reach the values suggested by some authors.9 On the other hand, patients with NMD were younger and had a much lower BMI. The fact that many of these patients had already received HNIV can explain why this group showed closer to normal PaCO2 values.
Also, it was possible to identify a heterogeneous group mainly composed of obese subjects with alveolar hypoventilation, most of which hadn’t been diagnosed before admission and were charac terized by high BMI, chronic respiratory acidosis (high level of blood bicarbonate) and acidemia (exacerbation).
The 37 survivors with NIV showed minimum pH deviation on admission (7.36 vs. 7.30) and a PaCO2 that was lower than that of patients who died (51 mmHg vs. 68 mmHg), thus indicating that the latter had severe respiratory failure. Besides, most patients in this sample were using home NIV; this means they continued using a ventilatory support that was familiar to them. More than 80% of patients in this group showed severe acidemia (pH < 7.35), so it would have been questionable to use limited resources from the ICU environment. On this regard, it is interesting to see that the existence of a non-invasive ventilation program avoided the use of valuable, scarce, and expen sive resources (intensive care bed) for patients with HNIV, many of which were hospitalized for intercurrent conditions and required control and supplementary monitoring exams for the purpose of optimizing the ventilatory support.
Almost half of the patients used level I devices (basic). Vital support devices were used in NMD patients who had greater dependence or needed to have multiple ventilation modes.5 Even though the PaCO2 decreased with NIV and was close to normal ity upon discharge (46 mmHg), not every patient resolved hypercapnia (10 patients were discharged with PaCO2 > 45 mmHg). We must mention that our institutional protocol involves follow-up through a daytime hospital model that prioritizes discharge with ventilatory support once the patient has adapt ed to ventilation and obtained basic comfortable parameters and improvement in pH and PaCO2.
Our care model is similar to the one used in North America and some European countries; the kinesiologist is included in the team that is responsible for delivering NIV. Furthermore, the nursing staff can detect and resolve problems and intolerance.19
According to the data reported by developed countries, one fifth of NIV treatments begin in a regular ward.20 However, logistic difficulties restrict the use of these treatments. In Europe, Australia and New Zealand, shortage of personnel and in frastructure limit the use of NIV treatments.21-25 A study of 157 centers in 51 countries of the five continents showed that 66% of them use NIV outside the ICU. Inadequate training and limited human resources are the reasons why NIV outside the ICU wasn’t implemented.26
In Latin America, data are limited. A survey conducted in fifteen hospitals in San Pablo, Brazil, showed that private hospitals made greater use of NIV, where kinesiologists seemed to be more skilled (100%) than physicians (73%) and nurses (33%).27
In Argentina, information is scarce. According to Alonso et al, there are some areas where NIV is not actively used28; and this could be considered an indicator of poor sanitary quality. Also, the or ganizational model of each center determines the use of NIV outside the ICU.29-31 Our institution has a daytime hospital where we begin HNIV;29 this can explain the differences with other series.15,18-20
Many patients with advanced cancer or progres sive diseases aren’t candidates for endotracheal intubation or invasive ventilation. A European study evaluated the acceptability and efficacy of NIV vs. conventional oxygen therapy to reduce dyspnea and the opioid dose. The results suggested that NIV is effective and comfortable for patients with terminal cancer.10
In our series, two thirds of patients with NIV and palliative care survived and were discharged. Deaths occurred in patients with advance direc tives for limitation of therapeutic effort. This finding shows the complexity of decisions at the end of life in cases of respiratory failure, the dif ficulty to predict the outcome and the role of NIV as “ceiling of treatment” according to Roberts et al, in a multicenter study of real-world.32
Azoulay et al conducted a prospective, multi center study on the use of NIV in terminal pa tients11 in 54 centers of France and Belgium. 134 patients with a “do not intubate” order, survivors on day 90, didn’t show a reduction in their quality of life. In our series, 43% of patients received NIV concurrently with opioids and anti-axiety drugs. However, this scenario is limited to specialized centers after a case-by-case discussion about the scope of treatment. In any case, NIV shall be con tinued only if it is well-tolerated by the patient and if a benefit is obtained. On the other hand, other measures must be taken to treat dyspnea (for example, drugs). It is necessary to take into account the fact that in some situations, NIV can unnecessarily prolong the life of the patient.33
This study has many limitations, included those inherent to retrospective studies, and since it is a single-center study with a heterogeneous popula tion, comparisons are difficult to make. There isn’t any control group, either, and most patients with severe exacerbation criteria had advance directives of limitation of therapeutic effort.
NIV was used in the regular ward environment, especially in patients with COPD, neuromuscular diseases and obesity with hypoventilation, and most of them had used some type of ventilatory support before being admitted and were discharged with indication of home ventilation. However, this per spective enlightens us about the use of resources in this specific population in a real-life scenario.
Both the number of treatment failures requiring ICU admission and the inpatient mortality rate were low. Deaths were recorded in patients with advance directives for limitation of therapeutic effort.