Respiratory Infections in Adults and Inequality: An Analysis of Deaths and Their Socioeconomic Determinants in Brazil

Abstract

Introduction: Respiratory infections cause serious complications responsible for a significant number of deaths in Brazil. In addition, the causes of death can be influenced by social and economic inequalities in Brazilian regions. Objective: To analyze the epidemiological profile and the influence of demographic and socioeconomic factors on deaths from respiratory infections in the adult population between 2014 and 2023 in Brazil. Methods: This was an analytical ecological study using data from the Death Information System. Death incidences were calculated. Multinomial logistic regressions and correlation tests were used to analyze the influence of socioeconomic factors on deaths. Results: There were high incidences of deaths from unspecified pneumonia, unconfirmed tuberculosis and complicated influenza. Deaths from pneumonia and the Gini index were positively correlated, considering the variables black ethnicity (R = 0.894), age over 90 (R = 0.869) and no schooling (R = 0.818) before the pandemic. The odds ratio of death from tuberculosis and influenza in the 70–79 age group (OR = 3.97) and black ethnicity (OR = 1.24), respectively, were higher in the pandemic and post-pandemic periods compared to the previous period. Conclusions: Deaths from respiratory infections were mainly influenced by demographic variables and socioeconomic inequalities in Brazil.

1. Introduction

The Unified Health System (SUS) is a public system of health services managed by the Brazilian Ministry of Health and is distributed across the political-administrative regions of Brazilian territory, such as the South, Southeast, Midwest, North and Northeast [1]. The SUS was developed with the aim of guaranteeing quality healthcare regardless of the nationality, socio-economic level or ethnicity of any patient. This contributes to the access of vulnerable populations to Brazilian health services [1].

However, the effectiveness of healthcare in the SUS is hampered by socio-economic disparities in Brazil’s regions [2]. As a result, some populations in these regions live in areas with precarious conditions of basic sanitation, work and health services and in places with high exposure to pollution generated by industries and fires [2].

These unfavorable conditions facilitate the spread of and deaths from respiratory infections in Brazil. According to the Ministry of Health, 552,490 deaths from respiratory diseases were recorded, of which around 58% were caused by respiratory tract infections, between 2015 and 2020 in Brazil [3].

Respiratory tract infections can be caused by microorganisms from different biological kingdoms, such as bacteria, viruses, fungi and parasites [4]. Cough, dyspnea, signs of respiratory effort, adventitious respiratory sounds, fever and malaise are possible clinical manifestations of these diseases [4].

Among these infections is tuberculosis, which is caused by Mycobacterium tuberculosis and can cause serious complications, including hemoptysis, pneumothorax, bronchiectasis, lung gangrene, fistulas and tracheobronchial stenosis [5]. This infection can also develop conditions susceptible to secondary fungal infections, such as chronic pulmonary aspergillosis [5]. The main method of preventing and reducing morbidity and mortality from tuberculosis is antibiotic therapy, which generates a substantial cost to the Brazilian health system, amounting to USD 20,800,598.40 [5]. This cost can be increased due to difficulties related to adherence to treatment, the long duration of therapy, the high number of antibiotics and the risk of adverse reactions [5].

Pneumonia is another major respiratory infection that can lead to fatal complications such as severe acute respiratory syndrome (SARS). These complications are often aggravated by comorbidities, postoperative conditions, transport accidents, and polytrauma, in addition to antimicrobial resistance mechanisms developed by pathogens such as Haemophilus influenzae, Klebsiella pneumoniae, Staphylococcus aureus, and Streptococcus pneumoniae [6].

In addition, pneumonia can develop as a secondary complication of viral infections, such as COVID-19, which was first identified in China in the final months of 2019. SARS-CoV-2 has spread rapidly to several countries, resulting in a high number of hospitalizations and deaths [7]. To reduce the spread of this virus, non-pharmaceutical interventions have been implemented [7]. During this period, it is likely that incidence and mortality from influenza and other acute upper and lower respiratory tract infections have been influenced by the reallocation of health service resources to deal with the COVID-19 pandemic [7].

Given the public health impacts of respiratory infections and the influence of various factors on mortality, this study aims to analyze the epidemiological profile and the influence of demographic and socioeconomic factors on deaths caused by respiratory infections in the adult population in Brazil from 2014 to 2023.

2. Materials and Methods

2.1. Study Design

This is an epidemiological, observational, ecological and analytical study of deaths caused by respiratory tract infections in adults during the years 2014 to 2023 in Brazil.

2.2. Data Extraction and Databases

In this study, the adult population was selected for the analysis of deaths due to the presence of different etiological agents, respiratory comorbidities, occupational exposures, and behavioral practices like smoking and alcoholism, in comparison to the pediatric population.

The information used in this study was exported from the databases of the Department of Information and Informatics of the Unified Health System (DATASUS) of the Ministry of Health. These databases provide important data for the development of strategies and interventions related to epidemiological surveillance and healthcare [8,9]. The data was extracted for each year of the period analyzed in this study.

(I)

Mortality Information System (SIM)

In Brazil, the death certificate (CD) must be filled out by the doctor with information on the underlying cause and secondary cause of death [10]. These documents are sent to the Municipal Health Departments and Civil Registry Offices, which code the causes of death according to the 10th International Classification of Diseases and Related Health Problems (ICD) [10].

The CD data is reviewed and improved to reduce errors, inconsistencies and duplicates [11]. After these adjustments, the data is recorded in the Mortality Information System (SIM) database [11], which makes this information available in .DBC files. SIM data is updated at the end of the year by adding data from the previous year [11].

In this study, these files were exported from the DATASUS File Transfer Platform on 31 December 2024. The .DBC files are not designed for direct import into Microsoft Excel. The .DBC files were then converted into .DBF files using the TABWIN software (http://siab.datasus.gov.br/DATASUS/index.php?area=060805&item=3, accessed on 5 May 2025), which is a DATASUS tool used to tabulate the data and create maps. The .DBF files were then imported into Microsoft Excel.

Afterward, these files are imported and converted into Excel spreadsheets, which consist of individual data obtained from the CD. Each row in these spreadsheets presents data from one death certificate. In this study, the following data were used: region, gender, ethnicity, age group, level of education, place of death, medical care performed, autopsy performed, and underlying cause of death coded using the International Classification of Diseases.

(II)

National Register of Health Establishments (CNES)

The CNES is a database that stores information on the number of health professionals and beds in public and private health establishments in Brazil [12]. CNES data is updated monthly by importing information from the previous month [12].

In this study, these CNES data were exported on 11 February 2025. The data were exported via spreadsheets that present aggregated data on the total number of healthcare professionals in the Brazilian public and private health systems, considering the month and year analyzed.

During the selection of data for export from the CNES website on DATASUS, doctors, nursing technicians and nursing assistants and beds available in outpatient clinics and in all hospitals in the Brazilian public and private health system were selected. No data was extracted for pediatric beds.

(III)

Population Projections

This study also used the 2024 population projections made available by the Brazilian Institute of Geography and Statistics (IBGE) [13], the Brazilian institution responsible for providing geographical, demographic and statistical information on the Brazilian territory [13]. This data was exported on 18 February 2025 in this study.

According to the IBGE, the population projections were calculated by incorporating and correcting the information obtained in the 2022 Demographic Census, which presented inconsistencies related to the demographic data collection process [14]. The files exported from the IBGE website provided aggregated data on population projections, considering sex, age, federative regions, and total numbers for Brazilian territory.

(IV)

Socio-economic indicators

The annual values for Brazil’s Gini index and Human Development Index (HDI) were obtained from the atlas and documents provided by the United Nations Development Programme (UNDP). In this study, the general coefficients of the Gini index and the HDI were used, which measure all types of inequality. Specific aspects or regional Brazilian variations of these indicators were not analyzed due to insufficient information on this topic [15,16]. This data was researched on 18 February 2025 in this study.

In this study, all the information and databases used are in the public domain and do not contain data that identifies the participants. This is in accordance with Resolution 466/2012 of the National Health Council, which is responsible for establishing standards and guidelines involving research with human beings in Brazil [16]. For this reason, this study did not require approval from the Research Ethics Committee.

2.3. Participants

The sample consisted of the total number of deaths from respiratory infections in the period analyzed. This included data from death certificates that were coded using the International Classification of Diseases (ICDs) for respiratory tract infections (

Table 1. Description of the classifications of deaths from respiratory infections based on the tenth International Classification of Diseases.

Respiratory System Infections	Description
Pneumonia
Unknown etiology	Infectious agents that have caused deaths from pneumonia, which may or may not have been identified through complementary tests. The bacterial category includes S. pneumoniae, H. influenzae and other bacteria. The viral category includes adenovirus, respiratory syncytial virus and others. The last classification includes deaths caused by P. jirovecii.
Pneumonia, unspecified (J18.9)
Bronchopneumonia, unspecified (J18.0)
Others (J18.1, J18.2 and J18.8)
Bacterial etiology
Bacterial pneumonia, unspecified (J15.9)
Specified etiologic agents (J13, J14, J15.0–J15.7)
Others (B01.2, B05.2, B25.0, J15.8 and P23.9)
Viral etiology
Viral pneumonia, unspecified (J12.9)
Others (J12.0–J12.8)
Others pneumonia
HIV disease resulting in Pneumocystis jirovecii pneumonia (B20.6)
Tuberculosis
Unconfirmed pulmonary (A16.0–A16.2)	Deaths caused by respiratory forms of tuberculosis, with or without bacteriological and histological confirmation.
Confirmed pulmonary (A15.0–A15.3)
Other respiratory forms of tuberculosis, without bacteriological or histological confirmation (A16.3–16.9)
Other respiratory forms of tuberculosis with bacteriological and histological confirmation (A15.4–A15.9)
Sequelae of tuberculosis of the respiratory tract (B90.9)
Influenza flu
Complicated with pneumonia (J10.0 and J11.0)	Deaths from flu with or without complications. Includes deaths with or without an infectious agent identified by complementary tests.
Other symptoms (J10.1, J10.8, J11.1 and J11.8)
Avian influenza virus (J09)
Other ICD
Acute lower airway infections (J21–J22)	Respiratory diseases with low numbers of deaths from infectious agents, not included in the other groups analyzed.
Acute upper airway infections (J00–J06)
Other pulmonary mycobacteria (A31.0)
Other pulmonary mycoses (B37.1, B80.0–B80.1, B39.0–B39.1, B40.0–B40.7, B41.0, B44.0–B44.1, B45.0, B46.0)
Others (B33.4 and B58.3)

).

However, deaths that were recorded by international disease classifications that unified abscesses and granulomas in the same category were not considered, as some granuloma deaths were caused by non-infectious diseases, nor were ICDs that recorded clinical conditions that indirectly predispose to death from infections.

The inclusion criteria were deaths with a record of sex and age greater than or equal to 18 years. A total of 1129 deaths that did not have this information recorded were excluded.

2.4. Procedures and Data Analysis

Jamovi 2.6.13 software and Microsoft Excel 365 were used for statistical analysis of the data. JAMOVI is a free platform that enables statistical analysis, and the integration of data used in other statistical programs [17]. A 95% confidence interval and 5% significance level were used for all analysis.

The pandemic period was defined as the interval between the years 2020 and 2022 according to the government decrees that considered the beginning and end of the health emergency by COVID-19 in Brazil [18]. The interval between 2014 and 2019 was considered pre-pandemic [18]. In this study, the year 2023 was considered a post-pandemic period due to the relaxation of social isolation measures having occurred at different times in the Brazilian states in 2022 [18].

Data analysis consisted of the following steps:

• The death data were aggregated according to the selected variables from the CD, using Excel’s “pivot table” feature. This way, the total numbers of deaths were obtained, considering each of the variables selected for analysis. The numbers of deaths were organized according to the variables on the death certificates and the infection groups. Deaths of individuals with complete or incomplete primary education, complete or incomplete tertiary education and deaths in hospitals and other health establishments were unified. Data that did not show a record of any of the variables on the death certificate were considered “not informed”. In Brazil, the mixed-race ethnicity does not include people of Asian, indigenous or other ethnicities.
• The total annual number of deaths was calculated for each group of infections and for each CD variable, such as region, sex, age group, ethnicity, level of educational attainment, location, medical care and autopsy.
• The relative frequencies of deaths were calculated for each infection group.
• The annual incidence of deaths was calculated by dividing the total number of deaths by the number of inhabitants. The result was then multiplied by one hundred. The incidences were calculated for each variable analyzed.
• The average annual incidence of deaths was calculated for pre-pandemic, pandemic and post-pandemic periods.
• The percentages of variation in the annual incidence of deaths for each period were calculated.
• Multinomial logistic regressions were used to analyze the odds ratios of deaths in the periods analyzed. For these tests, 423,682 death records classified as unknown for each death certificate variable were excluded. Nagelkerke’s pseudo R² was used and the enter method was employed. The periods (pre-pandemic, pandemic and post-pandemic) and CD variables were considered the dependent variable and the factors, respectively. Each research period (pre-pandemic, pandemic, and post-pandemic) was considered a dependent variable. The death certificate variables were selected as death certificate factors. These variables were organized into groups: sex, age group, education level, place of death, medical assistance, and autopsy on the death certificate. The odds ratio analyzes the relationship of each variable with the reference variable within its respective group. For the analysis of odds ratios, the following variables were considered as references: pre-pandemic and pandemic period, gender, 18–19 age group, white ethnicity, illiteracy, hospitals, not having received medical assistance before death and deaths in which no autopsies were carried out. Relationships between the pre-pandemic and post-pandemic periods were not analyzed in this study due to the focus on examining the association of successive periods.
• The average annual numbers of healthcare professionals and beds were calculated by averaging the total monthly numbers of healthcare workers and beds, respectively.
• The Shapiro−Wilk test was used to analyze the normality of the total number of deaths, The Human Development Index (HDI) and Gini coefficient values and the average annual number of healthcare workers and beds in the pre-pandemic and pandemic periods. This test was not used in 2023 because there are unique numbers for each variable.
• To analyze the correlation between the total number of deaths from respiratory infections and the average annual numbers of healthcare professionals and beds and The Human Development Index (HDI) and Gini coefficient values. Pearson and Spearman tests were used. Pearson’s correlation test was used for variables with a normal distribution. Spearman’s correlation test was used for numbers with a non-parametric distribution. Pearson’s correlation test was used for variables with a normal distribution. Spearman’s correlation test was used for numbers with a non-parametric distribution. Correlations were classified as positive or negative. Correlations with Pearson or Spearman coefficients greater than 0.7 were considered strong. Correlations with Pearson or Spearman coefficients equal to 1 were considered perfect.

3. Results

In Brazil, 14,179,520 deaths were recorded, of which 2,283,825 (16.10%) were caused by infectious diseases and 5.66% were caused by respiratory infections in the adult population during the period from 2014 to 2023. In addition, respiratory infections account for 35.13% of deaths from infectious diseases in adults.

During the years analyzed, there was a predominance of deaths from respiratory infections caused by unspecified agents, mainly pneumonia (

Table 2. Incidence of deaths from respiratory infections in the pre-pandemic (PRE), pandemic (PAN), and post-pandemic (POS) periods in Brazil. Percentage variations in the incidence of deaths in the pandemic period compared to the pre-pandemic period (%PRE) and in the incidence of deaths in the post-pandemic period compared to the pandemic period (%PAN) in Brazil. Relative frequency (RF) of total deaths from infections, from 2014 to 2023.

Groups	PRE	PAN	%PRE	POST	%PAN	Total	RF
Pneumonia	49.73	44.37	−10.78	50.24	13.22	736,489	91.79
Unknown etiology	41.87	30.46	−27.27	38.09	25.07	580,403	72.34
Pneumonia, unspecified	32.03	23.88	−25.46	30.09	26.02	448,276	55.87
Bronchopneumonia, unspecified	8.75	5.70	−34.88	6.77	18.91	116,171	14.48
Others	1.09	0.88	−19.28	1.23	39.29	15,956	1.99
Bacterial etiology	7.26	11.79	62.34	11.46	−2.8	139,665	17.41
Bacterial pneumonia, unspecified	6.35	10.58	66.64	10.25	−3.14	123,682	15.42
Other pneumonias	0.65	0.94	44.26	0.88	−5.97	11,717	1.46
Specified etiologic agents	0.27	0.28	4.09	0.34	20.79	4266	0.53
Viral etiology	0.18	1.86	950.63	0.38	−79.38	10,982	1.37
Viral pneumonia, unspecified	0.12	1.62	1199.51	0.26	−84.13	9169	1.14
Others	0.05	0.24	358.97	0.13	−47.38	1813	0.23
HIV disease resulting in Pneumocystis jirovecii pneumonia	0.42	0.26	−38.15	0.30	14.66	5439	0.68
Tuberculosis	2.66	2.90	9.19	3.27	12.78	42,774	5.33
Unconfirmed pulmonary	1.62	1.48	−9.17	1.60	8.11	24,064	3.00
Confirmed pulmonary	0.42	0.68	60.89	0.88	30.27	8418	1.05
Other unconfirmed respiratory forms	0.33	0.41	21.70	0.40	−2.71	5561	0.69
Other confirmed respiratory forms	0.08	0.20	134.61	0.26	31.13	2119	0.26
Sequelae of respiratory tuberculosis	0.19	0.14	−26.04	0.14	−3.34	2612	0.33
Influenza flu	0.53	1.09	104.55	0.69	−36.56	11,028	1.37
Complicated with pneumonia	0.21	0.46	117.12	0.30	−34.60	4574	0.57
Other symptoms	0.11	0.53	383.16	0.27	−48.71	3949	0.49
Avian influenza virus	0.21	0.09	−55.23	0.11	22.99	2505	0.31
Other ICD	0.59	1.04	76.29	1.14	9.84	12,054	1.50
Acute lower airway infections	0.29	0.54	84.63	0.59	10.08	6102	0.76
Acute upper airway infections	0.12	0.21	75.85	0.28	33.87	2515	0.31
Other mycobacteria	0.10	0.21	119.15	0.18	−14.63	2152	0.27
Other mycoses	0.07	0.07	4.46	0.08	7.38	1066	0.13
Others	0.02	0.01	−26.19	0.01	23.53	219	0.03
Brazil	53.51	49.40	−7.69	55.34	12.03	802,345

). During the pandemic, only the categories of viral and bacterial pneumonia, confirmed pulmonary tuberculosis and influenza complicated with pneumonia saw an increase in the incidence of deaths (Table 2).

However, there were increases in the incidence of pneumonia and reductions in the incidence of influenza in 2023 (Table 2). There was a progressive upward trend in the incidence of tuberculosis in all periods (Table 2).

During the pandemic, there was a reduction in the incidence of pneumonia, considering most of the variables, except for the increases obtained in the mixed race and black races and in the middle and higher education levels (

Table 3. Average incidence, percentage variations and total of deaths due to respiratory tract infections in the period from 2014 to 2023, in Brazil, considering the federative regions, sex, age group, race, schooling, place of occurrence, medical care and necropsy. Incidence of deaths from respiratory infections in the pre-pandemic (PRE), pandemic (PAN), and post-pandemic (POS) periods in Brazil. Percentage variations in the incidence of deaths in the pandemic period compared to the pre-pandemic period (%PRE) and in the incidence of deaths in the post-pandemic period compared to the pandemic period (%PAN) in Brazil.

Variables	2014		2015	2016	2017	2018	2019	2020	2021	2022	2023
Gini index	0.518		0.519	0.537	0.539	0.545	0.544	0.524	0.544	0.518	0.518
The Human Development Index	0.756		0.756	0.758	0.760	0.761	0.766	0.758	0.754	0.760	0.760
Average annual number of health professionals	1,586,129		1,652,944	1,731,749	1,828,759	1,953,864	2,074,475	2,246,904	2,521,958	2,689,827	2,868,082
Average annual number of beds	598,027		589,078	587,690	591,662	593,917	594,349	629,364	665,670	651,911	647,727
Periods	PRE	PAN	%PRE	POST	%PAN	Total	PRE	PAN	%PRE	POST	%PAN	Total
Groups	Pneumonia						Tuberculosis
Federal Regions
Southeast	65.44	54.82	−16.22	63.85	16.47	409,546	2.81	3.13	11.33	3.48	11.15	19,681
South	45.83	38.28	−16.48	42.49	11.00	98,155	2.04	2.35	15.13	2.64	12.35	5.01
Midwest	38.19	35.42	−7.24	41.27	16.52	44,208	1.76	2.28	29.39	2.53	10.80	2355
Northeast	34.72	35.33	1.77	38.71	9.55	143,115	2.69	2.60	−3.44	3.05	17.04	10,931
North	31.76	38.19	20.24	38.95	1.99	41,465	3.70	4.27	15.56	4.79	12.02	4797
Sex
Masculine	50.80	46.72	−8.04	52.15	11.62	365,166	4.13	4.61	11.67	5.16	11.92	32,206
Feminine	48.74	42.2	−13.42	48.47	14.86	371,323	1.30	1.32	2.06	1.53	15.63	10,568
Age Group’s
18–19	1.78	1.51	−15.34	1.70	12.56	1,12	0.29	0.50	71.02	0.65	30.44	253
20–29	2.63	2.27	−13.5	2.70	18.64	8412	0.72	1.00	37.95	1.19	19.56	2834
30–39	4.96	4.23	−14.76	4.65	10.05	15,647	1.30	1.63	25.60	1.83	12.50	4816
40–49	10.61	9.13	−13.99	9.68	6.05	28,473	2.46	2.62	6.23	3.20	22.21	7353
50–59	24.13	21.3	−11.73	22.13	3.92	53,065	3.95	3.96	0.45	4.16	5.07	9159
60–69	62.05	56.18	−9.46	56.81	1.12	94,363	5.27	5.26	−0.22	5.92	12.66	8452
70–79	199.75	165.67	−17.06	176.15	6.33	158,477	6.97	6.43	−7.75	6.60	2.73	5754
80–89	746.72	564.96	−24.34	625.25	10.67	229,422	10.36	9.47	−8.58	9.35	−1.32	3385
90 or more	2532.10	2031.78	−19.76	2438.83	20.03	147,510	13.18	11.08	−15.94	11.57	4.38	768
Ethnicity
White	29.73	25.05	−15.76	28.71	14.60	431,155	0.86	0.92	7.41	1.00	8.47	13,671
Mixed Race	14.46	14.62	1.11	16.53	13.05	225,546	1.29	1.43	10.48	1.66	16.23	20,991
Black	3.26	3.35	2.91	3.79	13.07	51,148	0.37	0.44	19.84	0.52	16.68	6.25
Others	0.50	0.45	−9.78	0.55	21.88	7514	0.04	0.04	5.36	0.04	7.79	589
Not informed	1.78	0.89	−49.72	0.66	−26.60	21,126	0.10	0.07	−31.1	0.05	−24.01	1273
Education
Illiteracy	10.16	8.60	−15.32	9.88	14.82	147,732	0.41	0.37	−10.35	0.41	11.81	6122
Primary	23.09	21.12	−8.54	24.85	17.65	346,949	1.38	1.54	11.12	1.76	14.59	22,496
Secondary	4.37	5.20	19.05	5.90	13.39	73,273	0.26	0.37	40.61	0.47	25.46	4865
Tertiary	1.92	2.34	22.23	2.71	15.89	32,623	0.06	0.08	38.7	0.10	22.73	1031
Not informed	10.20	7.11	−30.32	6.90	−2.90	135,912	0.54	0.54	0.77	0.53	−1.74	8.26
Place of Death
Health establishments	45.25	40.90	−9.61	46.27	13.13	673,484	2.23	2.44	9.57	2.80	14.87	35,983
Domicile	3.93	3.03	−22.97	3.44	13.72	55,055	0.37	0.40	9.40	0.40	−0.97	5854
Others	0.55	0.44	−20.22	0.52	18.90	7805	0.06	0.06	−6.99	0.07	24.31	920
Not informed	0.01	0.01	−27.2	0.00	−37.92	145	0.00	0.00	67.60	0.00	0.00	17
Healthcare
Yes	29.61	26.18	−11.58	29.98	14.54	437,416	1.50	1.63	8.89	1.95	19.51	24,295
No	1.38	1.18	−14.28	1.35	14.08	20,138	0.15	0.17	9.25	0.17	2.83	2463
Not informed	18.75	17.01	−9.27	18.9	11.13	278,935	1.00	1.10	9.63	1.15	4.32	16,016
Necropsy
Yes	3.05	1.16	−62.03	2.13	84.00	36.23	0.30	0.17	−42.67	0.29	66.12	3963
No	29.55	26.8	−9.29	30.16	12.52	440,151	1.48	1.73	17.2	1.96	13.21	24,584
Not informed	17.13	16.41	−4.23	17.95	9.37	260,108	0.88	1.00	13.43	1.03	2.83	14,227
Groups	Influenza						Others
Federal Regions
Southeast	0.48	1.06	120.35	0.51	−51.34	4375	0.47	0.69	46.07	0.72	4.61	3713
South	1.00	0.79	−21.43	0.89	13.51	2108	0.58	0.82	41.42	0.96	16.71	1583
Midwest	0.89	0.86	−3.07	0.9	3.79	1039	0.33	0.53	59.35	0.50	−4.78	481
Northeast	0.29	1.23	322.73	0.67	−45.77	2523	0.90	1.96	117.8	2.22	13.51	5529
North	0.38	1.52	302.18	1.10	−27.93	983	0.48	0.82	68.86	0.82	0.48	748
Sex
Masculine	0.57	1.08	89.7	0.63	−41.76	5386	0.65	1.15	76.87	1.30	13.40	6403
Feminine	0.50	1.09	120.22	0.75	−31.83	5642	0.54	0.94	75.67	1.00	5.85	5651
Age Group’s
18–19	0.09	0.07	−16.62	0.05	−34.3	53	0.09	0.10	15.77	0.11	13.77	61
20–29	0.08	0.12	55.05	0.11	−6.53	306	0.09	0.11	14.04	0.16	47.7	343
30–39	0.17	0.16	−6.32	0.16	−4.02	557	0.12	0.18	55.7	0.21	15.24	477
40–49	0.45	0.29	−34.12	0.29	−1.22	1087	0.25	0.39	58.72	0.43	10.67	889
50–59	0.72	0.64	−11.51	0.47	−25.81	1554	0.44	0.74	67.64	0.73	−0.18	1301
60–69	0.81	1.46	81.3	0.86	−40.9	1646	0.98	1.68	71.43	1.56	−6.88	2027
70–79	1.3	4.11	217.35	2.10	−49.05	1999	2.19	3.65	67.09	3.66	0.11	2439
80–89	3.61	11.83	227.37	6.66	−43.67	2276	6.23	10.57	69.62	11.36	7.46	2809
90 or more	13.79	42.99	211.81	25.41	−40.88	1.55	21.00	32.77	56.03	41.83	27.65	1708
Ethnicity
White	0.33	0.52	56.12	0.37	−28.42	6051	0.30	0.46	54.8	0.51	11.38	5656
Mixed Race	0.15	0.44	194.96	0.24	−45.04	3782	0.22	0.44	95.88	0.47	8.73	4821
Black	0.03	0.09	222.44	0.06	−36.52	782	0.05	0.09	98.32	0.11	26.00	1018
Others	0.01	0.01	86.87	0.01	−31.33	126	0.01	0.01	149.69	0.01	14.31	133
Not informed	0.02	0.03	72.32	0.01	−58.45	287	0.02	0.04	107.39	0.03	−31.58	426
Education
Illiteracy	0.08	0.24	190.23	0.15	−35.97	2092	0.12	0.20	68.79	0.23	11.65	2416
Primary	0.23	0.49	109.12	0.29	−41.55	4874	0.25	0.43	71.8	0.50	14.6	5124
Secondary	0.08	0.13	61.52	0.10	−23.58	1547	0.06	0.14	125.46	0.16	19.73	1449
Tertiary	0.04	0.05	26.58	0.05	6.18	702	0.02	0.06	133.91	0.07	16.28	606
Not informed	0.09	0.18	90.67	0.09	−45.98	1813	0.13	0.21	58.36	0.19	−10.16	2459
Place of Death
Health establishments	0.46	0.83	82.35	0.54	−35.86	8932	0.49	0.89	81.06	0.97	9.42	10,191
Domicile	0.07	0.24	257.11	0.14	−41.18	1933	0.09	0.13	52.11	0.16	16.07	1673
Others	0.01	0.02	124.09	0.01	−6.85	158	0.01	0.02	55.18	0.01	−17.47	188
Not informed	0.00	0.00	−38.70	0.00	197.31	5	0.00	0.00	84.79	0.00	0,00	2
Healthcare
Yes	0.38	0.61	62.27	0.44	−29.00	7025	0.36	0.59	61.53	0.64	8.77	7059
No	0.02	0.10	445.25	0.07	−33.53	770	0.03	0.04	27.14	0.04	18.62	504
Not informed	0.13	0.37	176.47	0.18	−50.05	3233	0.20	0.42	110.59	0.46	10.58	4491
Necropsy
Yes	0.05	0.03	−34.97	0.04	17.30	655	0.03	0.02	−28.38	0.03	40.42	416
No	0.36	0.70	93.60	0.47	−32.28	7339	0.37	0.64	71.09	0.68	6.25	7434
Not informed	0.12	0.35	194.91	0.18	−49.93	3034	0.19	0.38	103.14	0.44	14.11	4204

). The incidence of tuberculosis increased in the younger age groups (Table 3).

In 2023, there was an increase in deaths caused by respiratory infections, considering most of the variables in the death certificates, such as sex, age group, ethnicity, place of death and others. There was only a reduction in deaths from influenza among individuals who had received prior medical care (Table 3).

Increases in the incidence of death by respiratory infections during the pandemic were detected only in the Northeast and North regions. In 2023, the incidence of death by respiratory infections increased only in the Midwest region.

During the pre-pandemic period, there was a worsening of the Gini index, an improvement in the HDI, an increase in the average annual number of health professionals and a decrease in the average annual number of beds compared to 2014 (Table 3).

The Gini index showed strong positive correlations with deaths from pneumonia, considering age over 90, black ethnicity and mixed race and all levels of schooling (

Table 4. Pearson’s r correlation coefficients with statistical significance (p < 0.05), to analyze the correlations between the total number of deaths from respiratory infections and the Human Development Index (HDI), the Gini index and the average annual number of health professionals and beds.

Variables	Gini Index	HDI	Health Professionals	Beds
Pneumonia
90 years or older	0.869	0.881	0.918	-
Black	0.870	0.867	0.900	-
Mixed Race	0.894	0.848	0.903	-
Illiteracy	0.881	-	0.867	-
Primary	0.901	0.825	0.872	-
Secondary	0.863	0.913	0.931	-
Tertiary	0.874	0.914	0.929	-
No Health Care	0.954	-	0.887	-
Tuberculosis
40 to 49 years old	−0.881	-	-	-
Secondary	-	0.901	-	-
No Health Care	0.899	0.941	0.985	-
Performance of autopsy	−0.909	-	-	-
Influenza
Brazil	-	-	0.750	-
Masculine	-	-	0.730	-
Feminine	-	-	0.756	-
60 to 69 years old	-	-	0.737	-
70 to 79 years old	-	-	0.878	0.775
80 to 89 years old	-	-	0.872	0.755
90 years or older	-	-	0.832	0.709
Black	-	-	0.873	0.780
Mixed Race	-	-	0.853	0.736
Illiteracy	-	-	0.817	0.700
Primary Education	-	-	0.756	-
Health establishments	-	-	0.678	-
Domicile	-	-	0.881	0.788
No health care	-	-	0.920	0.871

). However, strong negative correlations were obtained between deaths from tuberculosis, considering the white race and the 40–49 age group (Table 4).

The HDI and physical resources showed strong positive correlations with the same variables correlated with the Gini index, except for deaths from pneumonia in the white population and individuals with no schooling (Table 4). Only deaths from tuberculosis in high school were correlated with the HDI (Table 4).

Human resources were positively correlated with deaths from influenza, considering both sexes, the elderly, non-white race, level of basic education and all places of occurrence (Table 4). In addition, physical resources were correlated with influenza deaths, considering most of the variables correlated with human resources (Table 4).

Autopsies performed on deaths from pneumonia and tuberculosis showed strong negative correlations with the Gini index and physical resources (Table 4).

No positive correlations were obtained between socio-economic factors and deaths from respiratory infections during the pandemic.

The odds ratios for death from pneumonia were higher in hospitals, in all ethnic groups and all levels of educational attainment, but deaths among female patients had lower odds ratios compared to deaths among male patients during the pandemic (

Table 5. Odds ratios (OR) obtained for deaths from respiratory infections using multinomial logistic regression tests. *The OR relates each variable from the death certificate to the reference variable within its respective group. The pre-pandemic and pandemic periods are the references for the analysis of odds ratios in relation to the pandemic and post-pandemic periods, respectively.

Periods		Pandemic vs. Pre-Pandemic						Post-Pandemic vs. X Pandemic
Groups		Pneumonia		Tuberculosis		Influenza		Pneumonia		Tuberculosis		Influenza
Nagelkerke pseudo-R-squared		0.79%		1.31%		11.60%		0.79%		1.31%		11.60%
CD variables	Reference	P	OR	P	OR	P	OR	P	OR	P	OR	P	OR
Sex
Feminine	Masculine	<0.001	0.96	0.013	0.91	0.047	1.12	0.053	1.02	0.301	1.06	0.054	1.19
Age Group’s
20–29	18–19	0.966	1.00	0.221	0.78	0.298	1.52	0.774	1.05	0.586	1.16	0.568	1.49
30–39		0.890	1.01	0.045	0.67	0.727	0.87	0.766	0.95	0.794	1.07	0.555	1.50
40–49		0.289	1.12	0.022	0.64	0.229	0.63	0.871	1.03	0.442	1.23	0.500	1.58
50–59		0.180	1.15	0.006	0.58	0.871	0.94	0.962	0.99	0.705	1.11	0.896	1.09
60–69		0.016	1.28	0.023	0.64	0.045	2.12	0.810	1.04	0.470	1.21	0.980	0.98
70–79		0.052	1.22	0.013	0.61	<0.001	3.97	0.598	1.09	0.596	1.16	0.749	0.81
80–89		0.339	1.10	0.014	0.61	<0.001	3.92	0.573	1.10	0.896	0.96	0.745	0.81
90 or more		0.133	1.16	0.065	0.66	<0.001	3.59	0.324	1.17	0.633	0.86	0.762	0.82
Ethnicity
Black	White	<0.001	1.25	0.056	1.10	<0.001	2.30	0.699	1.01	0.003	1.24	0.394	1.14
Mixed Race		<0.001	1.25	0.275	1.04	<0.001	2.04	0.619	1.01	0.059	1.10	<0.001	0.69
Education
Primary	Illiteracy	<0.001	1.13	<0.001	1.27	0.143	1.12	<0.001	1.06	0.857	0.99	0.457	0.92
Secondary		<0.001	1.45	<0.001	1.61	0.041	1.23	0.415	1.02	0.332	1.09	0.690	1.06
Tertiary		<0.001	1.46	<0.001	1.56	0.412	0.90	0.076	1.05	0.413	1.12	0.033	1.48
Place of Death
Domicile	Health establishments	<0.001	1.07	<0.001	1.38	0.282	1.11	<0.001	0.90	0.014	0.82	0.934	0.99
Healthcare
Yes	No	0.041	0.94	0.089	1.17	<0.001	0.34	<0.001	1.16	0.014	0.85	<0.001	2.93
Necropsy
Yes	No	<0.001	0.43	<0.001	0.45	<0.001	0.34	<0.001	1.73	<0.001	1.79	0.002	1.74

).

Compared to the pre-pandemic period, higher odds ratios were obtained for tuberculosis deaths at elementary school level and for flu deaths in the older age groups. During this period, deaths caused by all respiratory infections had lower odds ratios for autopsies and prior medical care (Table 5).

In 2023, higher odds ratios were obtained for deaths from pneumonia, influenza and tuberculosis in primary and higher education levels and in black ethnic groups, compared to the pandemic (Table 5). Mixed-race groups also had a higher odds ratio of death from influenza after the pandemic (Table 5)

In addition, the odds ratios for autopsies and prior medical care in deaths from pneumonia and influenza were higher in 2023. The odds ratios for deaths from these diseases at home were lower after the pandemic. (Table 5).

4. Discussion

This study found a predominance of deaths from unspecified pneumonia and a progressive increase in the incidence of deaths from tuberculosis over the ten years analyzed. During the pandemic years and in 2023, increases and decreases were detected in the incidence of deaths from pneumonia, respectively. The incidence of deaths from influenza, viral and bacterial pneumonia and other infections increased during the pandemic and decreased in 2023.

The incidence of deaths from tuberculosis increased during the pandemic, considering non-elderly age groups, health facilities, households and all levels of education. In addition, the incidence of death from influenza increased in the elderly population, all levels of education and places of death during this period.

An improvement in the human development index and a worsening of the Gini index were detected in the pre-pandemic years, a worsening of these indicators in the pandemic and a stabilization of the values of these indicators in 2023. During the pre-pandemic period, these indicators were positively correlated with deaths from pneumonia and tuberculosis, considering some death certificate variables. On the other hand, deaths from influenza were correlated with the average number of beds and health professionals during this period.

The odds ratios for death from respiratory infections were higher during the pandemic, considering low levels of education, non-white ethnic groups, failure to provide medical care and failure to perform autopsies. However, it should be noted that this study found unexpected results, such as higher odds ratios for deaths from infections among people of non-black ethnicities and people with a higher level of education in 2023.

4.1. Impact of Socio-Economic Factors

In this study, no correlations were found between deaths from influenza and the Gini index and the human development index (HDI). These findings may be related to negligence in caring for infectious diseases in Brazil and the high mutagenic variability of the influenza virus [19,20]. These factors can make it difficult to diagnose, prevent and treat influenza and can facilitate the spread of this disease, regardless of socioeconomic status as a social determinant of death [19,20]. On the other hand, deaths from influenza were correlated with the number of human and physical resources, which may be indirectly related to the worsening socioeconomic condition due to the structural and economic crisis of the Brazilian health system [21].

In contrast to the findings related to influenza, this study identified significant correlations that may be related to the impact of social inequality on deaths from other respiratory infections. The Gini index showed a worsening trend in the years analyzed and obtained positive correlations with the increase in deaths, considering various variables in the death certificate. This may indicate that growing economic inequality in Brazil increases vulnerability to infectious diseases [21].

Paradoxically, we detected a positive correlation between the improvement in the HDI and the increase in deaths from pneumonia and tuberculosis in the pre-pandemic period. This unexpected finding may be related to the fact that the overall improvement in the HDI in Brazil was mainly driven by a reduction in mortality from chronic non-infectious diseases [22,23]. Despite this, there is still neglect and a worsening of infectious diseases, which continue to disproportionately affect the most vulnerable populations [22,23].

In addition, this study found that deaths among black and brown people and people with low levels of schooling had higher odds ratios for death from respiratory infections. These findings are in line with international literature, which shows that minority populations with less access to education face barriers to preventive medicine and have a higher prevalence of comorbidities [24]. The greater susceptibility of non-white individuals was particularly pronounced during the pandemic, possibly due to greater exposure in jobs that did not allow for social isolation [25]. Similarly, the deaths of the elderly cannot only be related to biological factors but also to the lack of public policies for healthy ageing and the limitations of the health system in providing continuous care for this population, especially the sick of black ethnicity in poverty [26,27].

The period of the COVID-19 pandemic presented a different scenario in which, despite the worsening of all social and economic indicators, no statistical correlations were detected between deaths and socio-economic factors. The lack of correlation, however, does not invalidate the devastating impact of the crisis, such as the generalized overload of the health system and the exacerbation of vulnerabilities in all socioeconomic strata of the Brazilian population [21,26].

In addition, deaths among individuals with low levels of education may have been aggravated by the educational difficulties imposed by the pandemic, such as deficiencies in vocational training, a shortage of resources for remote education and a drop in the income of students from rural areas, who had to help their families with work [21]. Disparities also manifested themselves geographically, with the collapse of the health network in the North and Northeast regions, which have historically been unequal in terms of resources, with precarious hospital infrastructure and a shortage or unavailability of hospital oxygen, exemplifying the lethal impact of the combination of structural weaknesses and a health crisis [27,28].

In addition, deaths from infections in the North and Northeast regions can be influenced by failures in epidemiological surveillance, public policies developed based on data that is not scientifically supported and the unpreparedness of health services to deal with pandemics caused by infectious diseases [29].

In this study, higher odds ratios for death were detected in non-black ethnic groups and those with higher levels of education in the post-pandemic period. This finding may signal the long-term effects of the Brazilian economic crisis, such as increased unemployment and company bankruptcies, also impacting the deaths of previously less vulnerable groups [21].

4.2. Difficulties in Etiological Differential Diagnosis

The predominance of deaths caused by pneumonia with no specified etiology was also detected in a study carried out in the Brazilian city of Belo Horizonte, in which 55% of cases of respiratory infection were complicated by severe acute respiratory syndrome (SARS) caused by an unspecified etiological agent [30]. The widespread use of these classifications may be related to difficulties in the etiological diagnosis of respiratory infections influenced by the occurrence of similar clinical manifestations [31]. In this context, primary pulmonary tuberculosis can be confused with clinical conditions considered to be recurrent pneumonia.

The predominance of deaths in hospitals may be related to the inadequate filling in of underlying causes using the ICD for unspecified respiratory infections in the deaths of patients hospitalized for external causes, comorbidities and post-operative surgical complications [32]. In these situations, infections can be complications related to metabolic responses to trauma, diabetes, obesity, infections, malnutrition, prolonged hospitalization, corticotherapy and other factors [32]. As a result of these factors, it is also possible that an excessive number of deaths are recorded as respiratory infections in healthcare facilities, especially unspecified pneumonia.

In addition, the recording of the ICD for unspecified pneumonia may be related to the existence of one or more infectious agents in cases of coinfections, which can make differential diagnosis difficult and may encourage the use of unspecified ICDs [32]. In a study of patients hospitalized at Tongii Hospital in Shanghai, China, an increase in the number of patients with COVID-19 and influenza was detected [32]. Co-infections can be aggravating factors for death from respiratory infections due to worsening clinical conditions caused by immunological mechanisms that induce a cytokine storm [32].

On the other hand, the increase in the average incidence of deaths from viral and bacterial pneumonia detected in this study may be related to the underreporting of cases and deaths from COVID-19 (ICD 34.2) due to the recording of causes of death using other ICDs, such as SARS, unspecified pneumonia, unspecified respiratory failure, sepsis and ill-defined causes of death [30,33]. The absence of previous immunological mechanisms, immunoprophylaxis and effective treatments against the SARS-CoV-2 virus variants may be related to the higher probability of death from COVID-19 compared to other diseases at the beginning of the pandemic [30].

Some of the findings of this study, such as the reduction in the average incidence of unspecified pneumonia and the lower odds ratios for medical care, may be related to the implementation of social isolation, restrictions on access to health services and the redistribution of human and physical resources aimed at combating COVID-19 [34,35]. In addition, the lower odds ratio for autopsies and the higher odds ratio for home deaths may be related to the reduction in autopsy services to prevent COVID-19 transmission [34]. In this context, it is possible that clinical history data was used to determine the underlying cause of death instead of performing autopsies [34].

In addition, the increase in the average incidence of death detected in this study may be related to the increase in tuberculosis mortality rates and the rate of abandonment of treatment for this disease and may be related to the reduction in the number of laboratory and imaging tests aimed at detecting M. tuberculosis and other microorganisms [30,34,35]. However, the higher odds ratio of medical care before death from tuberculosis was an unexpected finding in this study, considering the implementation of coronavirus control measures. This finding may be related to patients seeking medical care for severe complications of tuberculosis, which increase the chance of death from this disease [30,34,35]. The increase in the average incidence of deaths from unspecified pneumonia and other infections detected in this study may be related to the discontinuation of the use of non-pharmacological measures and the redirection of economic resources to other pathologies, due to the reduction in mortality from COVID-19 [36].

Variations in the numbers, incidences and odds ratios of deaths from respiratory infections may be related to the lack of connection between the different government institutions that import this information into databases in Brazil [37]. In addition, this registration depends on manual work, which can lead to delays in importing deaths into information systems and reduce the reliability of the information [37].

4.3. Factors Influencing Susceptibility to Respiratory Infections

The reduction in the average incidence of deaths from unspecified pneumonia may be related to the reduction in the circulation of infectious agents due to the mandatory use of non-pharmacological measures to prevent and reduce the spread of the SARS-CoV-2 virus [38].

The spread of respiratory microorganisms can be reduced by filtering droplets and aerosols contaminated with pathogenic particles using face masks [38,39]. In addition, the application of soap and water or alcohol and gel to sanitize hands helps in the denaturation of proteins, contributing to the inactivation and death of respiratory infectious agents [38,39]. The effectiveness of non-pharmacological measures depends on the correct use of face masks and the combined application of these measures, which provide superior protection compared to the isolated application of each pharmacological intervention [38,39].

As in this study, reductions in deaths and cases were also observed in international studies a few months after the implementation of these measures, through a reduction in positive laboratory tests for respiratory agents in international surveys [36].

For example, in a survey conducted with pediatric participants in the United States of America from March to May 2020, reductions of approximately 98% were detected in the number of laboratory samples positive for the influenza virus, compared to the numbers obtained from September to February 2020 [40]. In addition, the number of laboratory samples positive for influenza was less than 0.2% in 2020 and 2021, compared to the number of positive materials, which were 2.36%, 1.04%, and 2.35% in 2019, 2018, and 2017, respectively [40]. A reduction in positive materials and lower odds ratios of laboratory samples positive for rhinovirus and enterovirus were detected compared to pre-pandemic years in another US study [41].

In addition, other studies have found that the decrease in the circulation of viral agents may have suppressed bacterial infections after the mandatory use of non-pharmaceutical measures began [42]. For example, in a study of children under 5 years of age in Israel, reductions in monthly rates of community-acquired bacterial and viral pneumonia were observed, concomitant with the total suppression of respiratory syncytial virus, influenza, and metapneumovirus circulation during the first phase of the pandemic [42].

On the other hand, this study detected increases in the average incidence of deaths from lower and upper respiratory tract infections and viral and bacterial pneumonia during the COVID-19 pandemic. Similar to these findings, another study detected significant increases in cases of rhinovirus and enterovirus in children under 5 years of age in emergency services and hospital departments during the period from October to February 2021 [39].

These increases may be related to the reduction in influenza cases due to influenza vaccination during the 2020–2021 influenza season, which contributes to the increase in cases of influenza-like syndromes caused by rhinoviruses, respiratory syncytial viruses, and other infectious agents [43].

In addition, these increases may be related to the characteristics of some non-enveloped viruses that can reduce the effectiveness of non-pharmacological interventions, such as virulence, prolonged elimination time of viral agents, resistance to simple disinfectants, and survival in extreme environmental conditions, such as gastrointestinal acidity, high temperatures, and dry climate [19].

This study showed increases in the average incidence of deaths from respiratory infections in the post-pandemic period, as did a study of children in Egypt, which recorded sudden and severe epidemics caused by respiratory syncytial virus in the pediatric and elderly populations a few months after the end of mandatory non-pharmaceutical interventions [44].

Although this study detected a reduction in the average incidence of deaths from bacterial pneumonia, a survey of children in England detected an increase in cases of infection by this virus that was simultaneous with the increase in cases of invasive pneumococcal disease in the second half of 2021, after the reduction in the mandatory use of non-pharmacological measures [45]. Another study conducted with patients of all ages in Germany recorded an increase in the incidence of invasive pneumococcal disease in the spring of 2021, reaching values similar to or higher than those in the years prior to the COVID-19 pandemic [46].

In the present study, an increase in the incidence of influenza deaths was detected during the COVID-19 health crisis, considering both sexes and the elderly. In addition, women had a higher odds ratio of death from influenza compared to men during the pandemic.

This increase in average incidence may be related to the ability of some influenza virus serotypes to block the spread of SARS-CoV-2 compared to respiratory syncytial virus and rhinovirus [47]. For this reason, some patients with influenza are less likely to be infected with coronavirus [47].

The lower odds ratio of death from non-influenza infections in women compared to men may be related to cultural factors that discourage them from seeking medical care [48]. These factors include the fear of showing weakness and the social representation of the male individual as the main provider of the family [48]. In addition, factors such as increased exposure to risky situations and the prioritization of work activities by men also contribute to their greater vulnerability and, consequently, to the increase in deaths in this group [48].

On the other hand, the average incidence of death from respiratory infections in the elderly and women may be related to the physiological decline that occurs in old age, pregnancy, puerperium, and climacteric [48]. During the pandemic, death may be related to decreased immune protection due to reduced repeated viral exposures resulting from the use of non-pharmacological measures [48].

Furthermore, the higher odds ratio of death from influenza may be related to the reduction in flu vaccination coverage among postpartum women, pregnant women, children, and the elderly in 2021, as detected in studies conducted in Brazil during the pandemic years [49]. These findings may have been influenced by the heterogeneous distribution of vaccines in Brazilian territory and outdated data on vaccination coverage in Brazil [48]. Additionally, the reduction in vaccination coverage may be related to the spread of false information about adverse reactions and the lack of vaccine safety [48].

In the present study, increases in the incidence of deaths and a higher odds ratio for death from tuberculosis were detected among young adults during the pandemic. These findings may be related to young adults’ greater involvement in practical work activities, which increases their contact with M. tuberculosis compared to the elderly [50]. Furthermore, the aging process and social isolation measures may have encouraged elderly individuals to remain at home [50].

In this study, high average incidences of deaths from respiratory infections were detected in the northern Brazilian region during the periods analyzed. These findings may be related to the humid equatorial climate of the Amazon rainforest, where the Northern region is located. A study conducted in Manaus, a Brazilian city in the state of Amazonas, showed correlations between weather patterns and the total number of hospitalizations for pneumonia in the months of April and May [51]. During these periods, high levels of rainfall and humidity were detected, while air temperature and the concentration of fine particulate matter decreased [51].

A direct correlation with rainfall and maximum air temperature and an inverse correlation with minimum air temperature were detected [52]. These correlations were detected mainly in the days before hospitalizations, when there were sudden changes in weather conditions [52]. These changes facilitate the prolonged suspension time of droplets and aerosols expelled by coughs and sneezes [51,52]. This favors the proliferation and spread of infectious agents and reduces immune defense mechanisms, contributing to death [51,52].

In addition to climatic factors, some environmental pollutants can prolong the suspension time and increase the hydrophobic composition, further favoring the spread of pathogenic microorganisms, especially in crowds of people in closed spaces with reduced ventilation [51,52]. Prolonged exposure to nitric dioxide and the oxidative effects of this pollutant contribute to weakened immunity and reduced lung function, contributing to death from respiratory diseases [52].

4.4. Limitations and Prospects

In the post-pandemic scenario (2023), the HDI and Gini index stabilized due to the existence of flaws in the recording of death certificates at the regional level, such as insufficient data, inappropriate recording of underlying causes of death and significant numbers of deaths due to unspecified etiological agents. In addition, a significant amount of uninformed data was detected regarding some death certificate variables and information regarding the post-pandemic period. These inconsistencies limited the proper counting of deaths and the analysis of the incidence of deaths and odds ratios.

The analysis of correlations between socio-economic indicators and physical and human resources and deaths from respiratory infections in the Brazilian regions was limited by the lack of data available from collection sources, such as documents from news media and the United Nations Development Program. In this context, no information was found on the specific and general measurements related to each socio-economic indicator at the level of Brazilian states and regions. In addition, the collection sources show variations in the values of these indicators, considering the collection source and the year in which the data was made available.

Factors related to the context of the pandemic have also limited the analysis of data from this study and international research, considering the targeting of government resources for the management of COVID-19 and the implementation of non-pharmacological treatment measures. From an international perspective, these adaptations may have affected the number of deaths by infection group, etiological agent and death certificate variables. In addition, the year 2023 also presented transition periods for the reduction in mandatory non-drug intervention measures, which increases the variability in the number of deaths recorded in the pandemic.

For further research, analyses could be carried out on deaths from infections caused by specified etiological agents confirmed by appropriate complementary tests. In this context, it is possible to include new variables, such as socioeconomic status, regional climatic factors, comorbidities and behavioral factors, such as smoking and alcohol consumption, which will allow us to improve the analysis of correlations with socioeconomic indicators. In addition, for future research it is possible to use specific measurements related to socio-economic indicators or to use new measurements that better represent the correlations between deaths and the socio-economic level of the patients.

5. Conclusions

This study analyzed the epidemiological profile of deaths from respiratory infections, showing a predominance of unspecified deaths, which may be related to difficulties in differential diagnosis, failure to record underlying causes of death and limitations of complementary tests. The analysis of the incidence of deaths, odds ratios and correlations, considering demographic and healthcare variables, reinforces the relationship between deaths and socioeconomic disparities among vulnerable groups.

Finally, it was observed that non-drug intervention measures may be related to the reduction in the number of deaths from some infections, and the worsening of social inequality in the pandemic context may be related to the higher odds ratio of death for vulnerable groups and groups that previously had better socioeconomic conditions. Despite the end of the pandemic, Brazil still has significant socio-economic problems and negligence in caring for respiratory infections.

This study contributes to avoiding neglect of respiratory infections, to encouraging the correct completion of death certificates and to encouraging measures to reduce the impact of these diseases.