Factores predictivos de fracaso de la ventilación mecánica no invasiva en la insuficiencia respiratoria aguda en paciente con neumonía por COVID19 ingresado en Unidad de Cuidados Intensivos

Abstract

Background The COVID-19 disease has posed a multidisciplinary challenge, especially for professionals in intensive care units. When approaching these critically ill patients, the choice of the type of ventilatory support plays a fundamental role. The use of non-invasive mechanical ventilation therapy offers advantages over invasive methods in managing these patients, mainly because it does not require anesthetic induction and continuous sedation as invasive methods do. Therefore, although Intensive Care Units (ICUs) have faced significant challenges, they have also achieved substantial progress thanks to the rapid generation of scientific evidence and its subsequent dissemination. This collaboration has facilitated the identification of effective treatments and therapies, the development of better practices for managing critically ill patients. Moreover, as the pandemic has evolved, control over it has improved, and ICUs have continued to adapt to the various changes caused by this global crisis. In this intensive medicine challenge to efficiently treat patients with severe acute respiratory syndrome due to COVID-19 who need ventilatory support, one of the primary aspects is to know the prognosis associated with the clinical picture to identify which factors influence the success or failure of such therapy. The main hypothesis is that the presence of certain risk factors, as well as the initial response in respiratory and clinical parameters, have an impact on patients and their prognosis. Specifically, patients have been characterized by having risk factors such as hypertension, cardiovascular diseases, diabetes, chronic respiratory diseases, or cancer. Regarding obesity, it has been identified as a severity factor that causes a more severe clinical course and greater complications. Specifically, it is associated with ventilatory dysfunction. Obesity, in turn, alters the balanced system of adipocytes and immune cells, consequently disrupting the immune system, which is also associated with greater complications and higher mortality. Therefore, determining which factors are associated with a worse prognosis is extremely important in the clinical performance of Intensive Care Medicine professionals as it can contribute to decision-making and more efficient resource redistribution. Similarly, determining which factors are associated with mechanical ventilation failure can allow us to anticipate the choice of the most suitable ventilatory support for each patient Objectives Therefore, in this retrospective cohort, the main objective has been to identify which predictive factors are associated with the failure of non-invasive mechanical ventilation therapy. Specifically, it has been of interest to analyze which factors, comorbidities, and superinfections were associated with higher mortality. Additionally, the impact of obesity, measured as BMI≥30, on the prognosis and mortality of these patients has been investigated, as well as the usefulness of the HACOR scale in predicting mortality and the failure of non-invasive mechanical ventilation (NIMV) during ICU stay. This aims to establish a predictive model of the success or failure of NIMV. Methodology: A retrospective multicenter cohort observational study was conducted, including 482 patients admitted for acute respiratory failure secondary to severe SARS-CoV-2 pneumonia in the ICUs of three national hospitals. Results/Discussion It has been identified that, at the initial ICU admission of patients in whom NIMV failed, there was a higher PaCO2 (46.3 vs. 35.9 mmHg, p<.001), FiO2 (100% vs. 65%, p<.001), heart rate (95 vs. 76 beats/minute, p<.001), and respiratory rate (32 vs. 28 breaths/minute, p=.065), as well as a lower pH (7.39 vs. 7.45, p<.001), diastolic blood pressure (64.5 vs. 71.1 mmHg, p=.018), and mean arterial pressure (87 vs. 93.9 mmHg, p=.048). Likewise, the SOFA score was higher in patients in whom NIMV failed (7 vs. 4, p<.001), as were some acute-phase reactants such as leukocytes (10.69 vs. 8.19 x 10³/μL, p=.001) and neutrophils (8.77 vs. 6.59 x 10³/μL, p<.001). The HACOR score was higher in patients in whom NIMV failed (8 vs. 6, p<.001). Other analytical variables, such as creatinine (0.70 vs. 0.84 mg/dL, p=.034), lactate (1.5 vs. 1.9 mmol/L, p=.057), and albumin (3.17 vs. 3.57 g/dL, p=.026), were lower in patients with NIMV failure. Subsequently, the variables directly related to NIMV failure with a p-value <0.1 were PaCO2 (OR 1.10), bicarbonate (OR 1.14), FiO2 (OR 1.05), heart rate (OR 1.04), respiratory rate (OR 1.04), leukocytes (OR 1.14), neutrophils (OR 1.16), LDH (OR 1.002), and SOFA score (OR 2.09). However, a history of chronic kidney disease (OR 0.14), MAP (OR 0.98), DAP (0.97), creatinine (OR 0.40), and albumin (OR 0.05) were inversely related (negatively) to NIMV failure. In the multivariate regression model, a direct relationship was identified with PaCO2 (OR 1.17), FiO2 (OR 1.04), heart rate (adjusted OR 1.04), leukocytes (OR 1.20), and SOFA score (OR 2.00), while MAP (OR 0.95) presented an inverse relationship with NIMV failure. Among the scales used to predict NIMV failure, the HACOR scale is the most widely used. Therefore, this thesis aimed to evaluate its usefulness in patients with COVID-19 disease admitted to ICU. A HACOR value >5 was used as a cutoff point to predict NIMV failure, obtaining a sensitivity of 84.2% (95% CI 60-96.6%), a specificity of 62.1% (95% CI 49.3-73.8%), a positive predictive value of 67.8% (95% CI 59.4-75.2%), and a negative predictive value of 80.6% (95% CI 59.1-92.3%). The coordinated function was applied to find the value of the variable that, when used as a cutoff point to discriminate between NIMV failure and success, maximizes sensitivity and specificity, obtaining a cutoff point of HACOR > 6.5. By modifying the cutoff point for the HACOR scale, specificity and positive predictive value improved, the latter to 72.9%. Another secondary objective of this study was to evaluate the clinical-epidemiological characteristics of patients with severe COVID-19 pneumonia admitted to ICU. A total of 480 patients were included; 335 men (69.5%) and 145 women, with a mean age of 61.94 ± 12.75 years. The most frequent comorbidities were hypertension (51.04%), obesity (23.44%), diabetes mellitus (23.44%), non-diabetic metabolic disease (21.16%), chronic heart failure (18.05%), COPD (11.62%), and chronic kidney disease (10.16%). Regarding mortality based on clinical-epidemiological characteristics, in this study, ICU-admitted COVID-19 pneumonia patients who survive are mostly men with an average age of 61.94 ± 12.75 years, which is considered high. Additionally, the most frequently reported comorbidities with the highest association with mortality are hypertension, obesity, diabetes mellitus, chronic heart failure, COPD, and chronic kidney disease. Multivariate analysis has shown that independent factors associated with higher mortality were the use of NIMV at ICU admission, the presence of diabetes mellitus as a pre-existing condition, and the time elapsed from hospital admission to NIMV initiation. Superinfections and other infectious complications in patients with SARS-CoV-2 pneumonia on NIMV play a crucial role in their prognosis, making it pertinent to analyze them by identifying the most frequent infectious foci, such as pulmonary (16.60%), bloodstream (7.68%), and genitourinary (7.05%). High-flow nasal oxygen therapy, being another widely used respiratory therapy in these patients, especially initially, has also been evaluated for its usefulness. Of the 327 patients with HFNO at ICU admission, 113 required orotracheal intubation (OTI) and connection to IMV. This represents an HFNO failure rate of 34.5%. A comparison was made between patients with HFNO who failed and those who did not fail during their total ICU stay, up to their discharge or death. In this regard, a higher need for FiO2 (70% vs. 60%, p<.001) was noted in patients in whom HFNO therapy failed, as well as lower PaO2 (63.1 vs. 73.3 mmHg, p<.001), SatO2 (91% vs. 94%, p<.001), and PaFi (99.3 vs. 128.7, p=.001), all compatible with a worse respiratory situation. Concurrently, these patients in whom HFNO failed had higher hospital mortality (39.8% vs. 8.9%, p<.001) and higher 90-day mortality (40.7% vs. 9.8%, p<.001). Finally, the influence of obesity, defined by a BMI >30, was examined, but in our patient selection, it was not associated with higher hospital mortality nor with higher mortality in the year following discharge. Similarly, obesity was not associated with a higher rate of NIMV failure or longer stay. Conclusions The increase in arterial CO2 pressure and heart rate, the need for high FiO2, the increase in blood leukocytes, and the SOFA scale score are directly related to an increase in the rate of NIMV failure, while the increase in mean arterial pressure acts as a protective factor against such failure. NIMV failure worsens prognosis and mortality. The predictors of mortality were the prior presence of diabetes mellitus, the time from hospital admission to IMV initiation, CRP levels at ICU admission, creatinine levels on day 3 in ICU, and PaFi values on day 3 in ICU. Prospective This doctoral thesis addresses essential issues that may influence the selection of ventilatory therapy in patients with COVID-19. By providing and identifying factors related to the failure of non-invasive ventilatory therapies, it may influence clinical decisions and improve the prognosis of these patients. The development of this predictive model would allow for the early detection of patients with a high probability of failure, enabling the redirection of resources more efficiently, equitably, and sustainably, as well as preventing complications and reducing mortality. In this regard, it may be appropriate to develop refined empirical studies that employ a randomized and controlled sampling methodology.