Factors related to mortality by pneumonia unrelated to mechanical ventilation

Objective: to analyze the factors related to mortality by nosocomial pneumonia unrelated to mechanical ventilation. Methods: retrospective cohort study with a sample acquired using the 538 notification forms for health care-related infections. The relative risk was calculated and a multivariable analysis was carried out using Poisson regression with a significance level of 5%. Results: the multivariable analysis showed that being under 59 years old and taking only one antimicrobial drug were protective factors against death. The main microorganisms responsible for the increase in the risk of death were: Acinetobacter baumannii, Klebsiella pneumoniae and species of Candida (Candida spp). Conclusion: the factors related to mortality were being 60 years old or older, using two or more antimicrobial substances, and being affected by the microorganisms Acinetobacter baumannii, Klebsiella pneumonia, and species of Candida (Candida spp). Descriptors: Pneumonia; Mortality; Cross Infection; Hospital Mortality. *Extraído da dissertação “Fatores de risco para óbito por pneumonia hospitalar não associada a ventilação mecânica”, Universidade Estadual de Londrina, 2019. 1Universidade Estadual de Londrina. Londrina, PR, Brasil. 2Instituto Federal do Paraná. Londrina, PR, Brasil. Autor correspondente: Carla Fernanda Tiroli Rua José Manoel Ruiz, 52, CEP: 86076-200. Londrina, PR, Brasil. E-mail: carla_tiroli@yahoo.com.br EDITOR CHEFE: Viviane Martins da Silva EDITOR ASSOCIADO: Renan Alves Silva Jucinay Phaedra Silva Sanches1 Carla Fernanda Tiroli1 Erika Bernardo da Silva1 Gabriela Machado Ezaias Paulino2 Gilselena Kerbauy1 Renata Aparecida Belei1 Flávia Meneguetti Pieri1 Como citar este artigo: Sanches JPS, Tiroli CF, Silva EB, Paulino GME, Kerbauy G, Belei RA, et al. Factors related to mortality by pneumonia unrelated to mechanical ventilation. Rev Rene. 2021;22:e62553. DOI: https://doi.org/10.15253/2175-6783.20212262553 Sanches JPS, Tiroli CF, Silva EB, Paulino GME, Kerbauy G, Belei RA, et al Rev Rene. 2021;22:e62553. 2 Introdução A pneumonia hospitalar adquirida, relacionada ou não com a ventilação mecânica, é uma das principais infecções presentes nos serviços de saúde e responsáveis por um elevado índice de mortalidade. De acordo com a Organização Mundial da Saúde, as infecções de vias respiratórias são altamente letais, causando três milhões de mortes em todo o mundo(1). O Brasil está entre os países com maior incidência dessa doença(2). Atualmente, é considerada um evento adverso persistente que ocasiona maior morbidade, aumento no tempo de internação e custos do tratamento, além da necessidade de cuidados intensivos com uso de ventilação mecânica e maior risco de mortalidade(3). Estudo multicêntrico norte-americano envolvendo 21 hospitais apontou uma taxa de incidência nos casos de pneumonia hospitalar não associada à ventilação mecânica de 0,12 a 2,28 casos por 1.000 pacientes/dia, dos quais 15,8% estiveram associadas ao óbito(4). Outro estudo em um hospital universitário, na região do nordeste descreveu que, do total das infecções relacionadas com a assistência à saúde, 30,2% foram diagnósticos de pneumonia(5). Vale destacar que foi considerada uma infecção subnotificada, uma vez que se observou a ausência de exigências para rastrear as taxas de incidência nos serviços(4,6-7). Em uma revisão narrativa, foi evidenciada uma lacuna científica na avaliação de medidas para prevenção e controle da pneumonia hospitalar não associada à ventilação mecânica. Sabe-se que os fatores de risco podem ser citados como não modificáveis, como idade, condições nutricionais e presença de comorbidades, e os modificáveis, relacionados com as condições da hospitalização, tratamentos fornecidos e contaminação ambiental(8-9). A pneumonia associada à ventilação mecânica foi e continua sendo estudada com grande interesse, especialmente, no que tange às medidas de prevenção e controle. No entanto, o conhecimento das medidas preventivas fez com que suas taxas de incidência reduzissem e a pneumonia não associada à ventilação mecânica ganhasse destaque, tornando-se uma das principais preocupações de segurança do paciente. Esta é considerada um evento adverso persistente nos serviços de saúde e, como consequência, ocasiona maior tempo de internação, além de morbimortalidade e aumentos nos custos do tratamento(9). Ressalta-se que o conhecimento sobre os fatores associados ao óbito permite o diagnóstico e tratamento precoce, fornecendo subsídios para o desenvolvimento de protocolos a respeito da prevenção e evolução clínica desses casos e auxiliando no melhor prognóstico ao paciente. Mediante o exposto, o presente estudo teve como objetivo analisar os fatores relacionados com mortalidade por pneumonia hospitalar não associada à ventilação mecânica.


Introduction
Acquired nosocomial pneumonia, related or not to mechanical ventilation, is one of the main infections in health services and is responsible for a high level of mortality. According to the World Health Organization, infections in the respiratory tract are highly lethal, causing three million deaths worldwide (1) . Brazil is one of the countries with the highest incidence of this type of disease (2) . Currently, it is seen as an adverse and persistent event, that leads to greater morbidity, increased hospitalization time and high treatment costs, not to mention the need for intensive care including mechanical ventilation, also leading to a higher risk of death (3) .
A North American multicentric study, involving 21 hospitals, presented a prevalence of nosocomial pneumonia unrelated to mechanical ventilation that varied from 0.12 to 2.28 cases per 1,000 patients/day. 15.8% of these cases were associated with death (4) . Another study in a university hospital, in the Brazilian Northeast, described that, from the total number of infections related to health care, 30.2% were diagnosed as having pneumonia (5) . It should be noted that this infection is considered to be undernotified, since there were no attempts to identify its rates of incidence in the services (4,(6)(7) .
A narrative review presented a scientific gap in the evaluation of measures to prevent and control nosocomial pneumonia unrelated to mechanical ventilation. Certain risk factors can be considered as unavoidable, including age, nutritional conditions, the presence of comorbidities, while others are avoidable, such as those related to conditions of hospitalization, to the treatment provided, and to environmental contamination (8)(9) .
Pneumonia related to mechanical ventilation was and still is investigated with great interest, especially regarding prevention and control measures. However, the knowledge about preventive measures led its incidence to diminish, and the pneumonia un-related to mechanical ventilation began to stood out, becoming one of the main preoccupations with the safety of the patient. It is considered to be a persisting adverse event in the health services and, as a consequence, it leads to longer hospitalization times, morbidity and mortality, and increases the cost of treatments (9) .
It should also be highlighted that knowing factors associated to death allows for early diagnosis and treatment, giving support to the development of protocols about the prevention and clinical evolution of these cases and helping to provide the patient with a better prognosis. Considering the above, this study aimed to analyze the factors related to mortality by nosocomial pneumonia unrelated to mechanical ventilation.

Methods
This study is a retrospective cohort. It was carried out in the university hospital of the Universidade Estadual de Londrina, in the city of Londrina, Paraná, Brazil.
The inclusion criteria were: patients with nosocomial pneumonia unrelated to mechanical ventilation, who were 18 years old or older; hospitalized in the emergency unit, hospitalization unit, intensive care unit, center for the treatment of burns, or in the intensive care unit for burns; and who were diagnosed with nosocomial pneumonia unrelated with mechanical ventilation, from January 2017 to December 2018. The diagnoses of nosocomial pneumonia unrelated to mechanical ventilation were carried out by the physician from the Commission for the Control of Nosocomial Infections, in accordance with the International Classification of Diseases (CID-10) and the definitions of the National Sanitary Surveillance Agency (8) . Were excluded notified cases whose forms had incomplete data. 540 patients were included, diagnosed with nosocomial pneumonia unrelated to mechanical ventilation. From these, 2 cases were excluded, because the data from the notification form was incomplete. As a result, the sample used in this research was comprised by 538 patients.
The information was acquired from a source of secondary data using the notification forms for health care related infections, made available by the Commission for the Control of Nosocomial Infections from the university hospital of the Universidade Estadual de Londrina.
In the hospital where this study was carried out, pneumonia cases are actively searched and identified, and are seen as an indication for the prescription of antimicrobial substances. These substances are only recommended after the approval from the Commission for the Control of Nosocomial Infections. Therefore, the work of this commission allows for the identification of all patients who were prescribed antimicrobial substances. After the patients with some type of infection were identified, a form of suspected health care-related infection is started. When the infection is confirmed, the records change into a form for the notification of health care-related infections.
The notification form was elaborated by the Commission for the Control of Nosocomial Infections according with the criteria from the National Agency of Health Surveillance, including clinical, laboratory, and imaging items. The form is filled in by undergraduate/interns who volunteer to participate in the commission, and are trained to do so by a nurse, who is PhD in education and responsible for the sector. This happens during the hospitalization of the patient, that is, the interns accompany and update information from notification to discharge or death.
All data from the form to notify nosocomial infections was collected and typed into an Excel spreadsheet. It should be mentioned that this form was already the one used for data collection and was considered to be complete by the team of the Commission for the Control of Nosocomial Infections, concerning data related to nosocomial infections.
The data collection stage took place from December 2019 to March 2020 by two researchers who were trained by the coordinator of the research project and by a nurse who had experience in the Commission for the Control of Nosocomial Infections and a member of the research project. Data collection took place once a week, in the afternoon, in the room of the commission. The following clinical and demographic variables were collected: gender, age, days of hospitalization, unit associated to the pneumonia, days of hospitalization until being diagnosed with pneumonia, blood culture associated with pneumonia, tracheal secretion culture associated to pneumonia, number of antimicrobial medicines used, and death.
The variables were categorized as follows: gender (male or female), age group (up to 59 years old or 60 years old or more), hospitalization days (up to 14 days or ≥15), unit associated to the pneumonia (emergency unit + hospitalization unit and intensive care unit + center for the treatment of burns + intensive care unit for burns), days of hospitalization until the diagnosis of pneumonia (up to 14 days or ≥15 ), blood culture and tracheal secretion culture associated with pneumonia (positive, negative, not collected) number of antimicrobial drugs used (1 or 2 or more), and death (yes or no).
The emergency unit was considered together with hospitalization units due to the profile of the institution, since patients from the emergency unit may not be transferred due to the lack of beds available. The center for the treatment of burns and the intensive care unit for the treatment of burns were grouped with the intensive care unit because these units receive severe cases until the clinical framework is stable.
The number of hospitalization days until diagnosis was calculated based on the subtraction between the variables: date of hospitalization, date when the nosocomial pneumonia unrelated to mechanical ventilation was diagnosed, and date of discharge or death. Diagnostic cultures were carried out upon request of the physician in charge of the patient, considering clinical aspects (hyperthermia, changes in the characteristics of trachea secretions), laboratory aspects (leukogram results), and imaging (x-ray).
Tracheal secretion collection is carried out via the aspiration of airwaves, when the patient is unable to promptly expel catarrh or was using catheters. For cases with no pulmonary secretion, no collection was made.
Blood culture was carried out through two venipunctures or peripheral artery punctures, and blood was stored in aerobic and anaerobic blood culture vials. After colection, the material was sent to the laboratory of clinical analysis of the university hospital from the Universidade Estadual de Londrina. There, tracheal secretion and blood culture were analyzed using the Bactec TM automated system. The absence of blood culture analysis is justified when the blood is collected by there is no microbial growth, when the team fails to collect the material and, finally, when the patient dies before there is time to collect a sample.
The data were transcribed into a Microsoft Excel ® 2013 spreadsheet. Later, they were transported and analyzed in the IBM software Statistical Package for the Social Sciences for Windows, version 20.0 ® For the bivariate analysis, the variable death was crossed with the following exposition variables: gender, age, unit associated to pneumonia, days of hospitalization prior to the diagnostic of pneumonia, blood culture, and tracheal secretion culture, associated to pneumonia, number of antimicrobial drugs used. The chi-squared test was applied, and the probability of type I error was considered to be significant when at 5%, with the respective confidence interval of 95% (CI 95%).
For the multivariable analysis, Poisson's regression with a robust variance adjustment was used. The selection of adjustment variables considered the following mathematical criteria: p<0.20 in the variate analysis, with the following variables for the multiva-riable model: age group, hospitalization days prior to the diagnosis of pneumonia, and number of antimicrobial drugs. The magnitude of the association was determined calculating relative risk (RR) with an CI of 95%. The standardized level of reference (RR=1) indicated the best situation possible, that is, no risk. The

Results
Little more than half participants were male (n=331; 61.5%) and 60 years old or older (n=327; 60.8%). Regarding clinical data, most were hospitalized for 15 days or more (n=406; 75.5%) and diagnosed with pneumonia for up to 14 days after hospitalization (n=451; 83.8%). The units associated to pneumonia the most were the emergency unit and hospitalization units (n=374; 69.5%). Most participants used one or more antimicrobial drugs (n=375; 69.7%) and 260 individuals (48.3%) died.
In the multivariable analysis, individuals who were 59 years old or younger (RR=0.60; IC95% = 0.50 -0.73; p<0.001) and who were applied only one antimicrobial drug (RR=0.35; IC 95% 0.26-0.47; p<0.001) were less likely to die when compared to those who were 60 years old or older and used more than one antimicrobial drug, respectively (Table 1). 398 patients had their tracheal secretion and/ or blood collected for bacterial cultures, and in 214 (53.8%) the culture was positive. Bivariate analysis of the bacterial cultures, associated with death, are presented in Table 2.  Patients with a positive culture associated with pneumonia had a higher risk of death (RR=1.51; IC95% = 1.24 -1.84; p<0.001). The increase in the risk of death was associated with positive cultures for the following microorganisms: Acinetobacter baumannii, Candida spp., and Klebsiella pneumoniae.

Discussion
The results of this investigation show the reality of a single local public hospital limiting the possibilities of generalizing the results, despite the sample size. It was also found that many forms are not adequately filled in and there is no indication of the severity of the cases. A limiting factor for this study is the fact that it used data from secondary sources, making it impossible to detect modifiable risk factors such as: environmental microbiota, assistance from the multiprofessional team, and unavoidable factors: score of severity at the time of admission and comorbidities.
The results presented here may give support to the care practice in the institution where the research was carried out through the elaboration of bundles, assistance protocols to detect nosocomial pneumonia unrelated to mechanical ventilation and treat it early, in addition to the development of later multicentric studies. Furthermore, it can contribute for teaching and research in nursing in regard with the acquisition of knowledge related to two factors that can provoke the death of patients with pneumonia unrelated to mechanical ventilation. It can also give support to a multifaceted management of care, which must be provided for the hospitalized patient to prevent health care-related infections, reflecting on the quality of attention and, consequently, on the reduction of the deaths associated to it.
The mortality in people with nosocomial pneumonia unrelated with mechanical ventilation is common, and the pneumonia can be a direct or indirect cause of their death. In a study carried out in the United States, the likelihood of death was significantly higher in people with nosocomial pneumonia unrelated to mechanical ventilation when compared to other gruops (10) .
The association between death and the number of antimicrobial drugs may be related to the resistance of the microorganisms present in the cultures, since the patient needs wide spectrum therapy and multiple antibiotics to fight against these pathogens. On the other hand, patients whose tests indicate antibiotic sensitivity and who are not in septic shock or have a high risk of death are recommended to undergo a single therapy (2,11) .
The main microorganisms associated to death, are: Acinetobacter baumanni, Klebsiella pneumoniae and Candida spp. The airways are naturally colonized, but when there are large aspirations and pathogen agents are absorbed, there is a risk for infection. This is especially true for elders with bad oral hygiene, bacterial biofilm, food residues, and erosion. The insertion of bacteria deteriorates the defense mechanisms, leading to the introduction of virulent pathogens (12) .
Acinetobacter baumanni pneumonia is an increasingly recognized infection in critical patients. High rates of resistance to the most common antimicrobial drugs mean that professionals must empirically choose the adequate treatment for pneumonia from Acinetobacter baumanni, which is challenging (13) . Similarly, infections by Klebsiella pneumoniae are difficult to treat due to the resistance against antimicrobial drugs and to hyper-virulent strains (14) .
Regarding Candida spp., a systematic review with meta-analysis showed that the colonization from this microorganism is associated to longer mechanical ventilation, more mortality in 28 days, hospitalization in intensive care units, and, probably, to longer hospitalizations (15) . The clinical, therapeutic, and demographic characteristics of the patients with nosocomial pneumonia unrelated to mechanical ventilation also show that the predominance of deaths in the elders can be attributed to the excess of comorbidities in this age group (16)(17) .
Although there have been many advances in this field regarding prevention, diagnosis and treat-ment for pneumonia associated to mechanical ventilation, prophylactic measures are insufficient when they are connected to nosocomial pneumonia unrelated to mechanical ventilation, leading to long hospitalizations due to the worsening of clinical conditions, and to the need of transferring into an intensive care unit and staying there (4,17) .
It is also worth reminding the causal relation with the hospitalization sector. Our results showed that the emergency unit and the hospitalization units were responsible for a greater incidence of cases. The precariousness of the care provided, caused by insufficient human resources, is one of the main causes of this health issue in hospital units, since, in intensive care units, the only well-defined protocols are targeted at preventing pneumonia related to mechanical ventilation (18) .
The risk of death by nosocomial pneumonia unrelated to mechanical ventilation, in addition to other components, must be investigated and analyzed, as a way to give more visibility to this issue. The knowledge and the ability of the health team and techniques associated to the prevention of infections, such as hand hygiene, are included. In this context, studies are needed that evaluate the assistance provided to the individual, the way in which oxygen therapy devices are stored in the units, the quality of chemical disinfection, not to mention studies that investigate the presence of microorganisms in the hospital environments (19) .

Conclusion
This research has found that the number of deaths in patients caused by nosocomial pneumonia unrelated to mechanical ventilation was high. The factors related to mortality were 60-year-old or older people, the use of two or more antimicrobial drugs, and the identification of the following etiologic agents: Acinetobacter baumannii, Klebsiella pneumoniae e Espécies de Candida (Candida spp.).

Collaborations
Sanches JPS, Tiroli CF, Silva EB and Pieri FM took part in the conception of the project, data analysis and interpretation, in the writing of the article, and in a relevant critical review of its intellectual content. Paulino GME, Kerbauy G, Belei RA took part in data interpretation, in a relevant critical review of the intellectual content, and in the approval of the final version to be published.