Factors Associated with the Occurrence of Severe Acute Respiratory Syndrome (SARS) among COVID-19 Cases

Abstract

This retrospective cohort study investigated factors associated with Severe Acute Respiratory Syndrome (SARS) or death among confirmed COVID-19 cases treated at a Primary Health Care (PHC) unit in Rio de Janeiro from August 2020 to January 2022. The occurrence of SARS or death was identified in the Influenza Epidemiological Surveillance Information System and in the Mortality Information System, respectively. Individual characteristics, symptoms, and predominant viral variant were the explanatory variables tested in logistic regression models. Of 1,708 COVID-19 episodes, 72 (4.2%) resulted in SARS, of which 17 (1.0%) were fatal. Among the risk factors for SARS, obesity, pregnancy, chronic heart disease, diabetes, dyspnea, fever, and illness during the predominance of the Gamma variant were highlighted. There were no cases of worsening during the predominance of the Omicron variant. In conclusion, the risk of COVID-19 worsening depends on individual patient characteristics as well as the circulating viral variants, underscoring the essential role of continuous surveillance in detecting changes in the epidemiological landscape and in guiding prevention and care strategies.

Key words:
COVID-19; Prognosis; Primary Health Care

Resumo

Este estudo de coorte retrospectivo investigou fatores associados à Síndrome Respiratória Aguda Grave (SRAG) ou óbito entre casos confirmados de COVID-19 atendidos em uma unidade de Atenção Primária à Saúde no Rio de Janeiro, entre agosto/2020 e janeiro/2022. A ocorrência de SRAG ou óbito foi identificada, respectivamente, nos Sistemas de Informação da Vigilância Epidemiológica da Gripe e de Informação sobre Mortalidade. Características individuais, sintomas e variante viral predominante foram as variáveis explicativas testadas em modelos de regressão logística. Dos 1.708 episódios de COVID-19, 72 (4,2%) resultaram em SRAG, dos quais 17 (1,0%) foram a óbito. Entre os fatores de risco para SRAG, destacaram-se obesidade, gravidez, doença cardíaca crônica, diabetes, dispneia, febre e adoecer no período de predomínio da variante Gamma. Não houve casos de agravamento durante o predomínio da variante Ômicron. Conclui-se que o risco de agravamento da COVID-19 depende de características individuais dos pacientes, bem como das variantes virais circulantes, enfatizando assim o papel essencial da vigilância contínua para identificar mudanças no cenário epidemiológico e orientar estratégias de prevenção e cuidado.

Palavras-chave:
COVID-19; Prognóstico; Atenção Primária à Saúde

Resumen

Este estudio de cohorte retrospectivo investigó factores asociados con el Síndrome Respiratorio Agudo Grave (SRAG) o la muerte entre casos confirmados de COVID-19 atendidos en una unidad de Atención Primaria de Salud en Río de Janeiro, entre agosto de 2020 y enero de 2022. La ocurrencia de SRAG o muerte fue identificada, respectivamente, en los Sistemas de Información de Vigilancia Epidemiológica de la Gripe y de Información sobre Mortalidad. Las características individuales, síntomas y variante viral predominante fueron las variables explicativas probadas en modelos de regresión logística. De los 1708 episodios de COVID-19, 72 (4,2%) resultaron en SRAG, de los cuales 17 (1,0%) fallecieron. Entre los factores de riesgo para SRAG, se destacaron obesidad, embarazo, enfermedad cardíaca crónica, diabetes, disnea, fiebre y haber enfermado durante el período de predominio de la variante Gamma. No hubo casos de agravamiento durante el predominio de la variante Ómicron. Se concluye que el riesgo de agravamiento de la COVID-19 depende de las características individuales de los pacientes, así como de las variantes virales circulantes, lo que resalta el papel esencial de la vigilancia continua para identificar cambios en el panorama epidemiológico y orientar las estrategias de prevención y atención.

Palabras clave:
COVID-19; Pronóstico; Atención Primaria de Salud

Introduction

COVID-19 commonly begins with flu-like symptoms, such as cough, fever, and body aches¹. Most cases have a self-limiting course. However, before vaccination began, approximately 15% of the patients developed extensive viral pneumonia, requiring hospitalization, and 5% required intensive care due to respiratory failure and dysfunctions in other organs, resulting in high mortality^2,3. In Brazil, by the end of 2021, the COVID-19 pandemic had caused more than 619,000 deaths⁴.

Identifying individuals with a high risk of developing severe disease is crucial for developing clinical protocols. Advanced age, being male, having diabetes mellitus, obesity, and immunosuppression were some of the first identified risk factors^5-8. A systematic review identified hundreds of models for predicting the risk of worsening of COVID-19, but most were conducted in hospital settings and had methodological limitations⁹. One population-based cohort study, conducted in the United Kingdom after the appearance of the Omicron variant, included more than one million cases and generated a prognostic score, the QCOVID4. The final prediction model incorporated age, sex, body mass index, ethnic group, deprivation score by area of residence, prior COVID-19, vaccination and a list with more than twenty health conditions⁶. However, specific sociodemographic characteristics included in the model prevent its use in other populations.

In Brazil, Primary Health Care (PHC) units are responsible for providing first aid to suspected cases of COVID-19 and diagnosing and monitoring mild cases. At this level of care, identifying individuals at a higher risk of worsening enables a more rational use of resources, guiding strategies, such as remote monitoring, home oximetry, and early antiviral therapy for the most vulnerable patients. In the United Kingdom, a predictive model for hospitalization was developed based on information from patients treated in PHC units between October 2020 and October 2021. The model, which included age, sex, hypertension, degree of dyspnea, and fever, was only suitable for predicting non-hospitalization in the lowest-risk group¹⁰.

Socioeconomic risk indicators are context-specific, and even clinical markers may differ between countries and regions with distinctive ethnic makeup and morbidity profiles. Moreover, throughout the pandemic, the emergence of viral variants and the immunity acquired by prior infection or by vaccination^11-13 resulted in changes in transmissibility, clinical presentation, and disease severity, adding more complexity to predicting the worsening of the disease.

The conduct of prognostic studies at PHC faces specific challenges, such as the rarity of severe events, which requires larger samples for robust analyses. In addition, until mid-2021, the low availability of diagnostic tests limited the confirmation of outpatient cases. Despite the high incidence and mortality of COVID-19, we identified a gap in the literature concerning studies focused on populations treated in PHC in Brazil.

The present study seeks to address this gap by evaluating risk factors associated with hospitalization and death among confirmed cases of COVID-19 treated at a PHC unit located in a low-income community in the city of Rio de Janeiro, thereby contributing to the development of strategies adapted to this context. The sociodemographic characteristics, symptoms, pre-existing health conditions, and vaccination status of the patients were evaluated, as was the viral variant predominating in each period throughout an 18-month period.

Methodology

Study design and location

This is a retrospective cohort study based on records of care provided at a PHC unit in the city of Rio de Janeiro, from August 26, 2020, to January 31, 2022. The occurrence of the outcome - Severe Acute Respiratory Syndrome (SARS) within 15 days or death within 30 days after care - was identified in the respective official databases.

The PHC unit serves an enrolled population of about 22,000 people and is located in an area with a Human Development Index (HDI) of 0.726, one of the lowest in the city. According to the service routine, all individuals who presented with fever or respiratory symptoms were evaluated sequentially by a nurse and a physician. In each consultation, a standardized form was filled out with closed fields for symptoms and health conditions considered as potential factors for the worsening of the disease^5,14. Individuals considered suspected COVID-19 cases¹⁴ were reported and sent for nasopharyngeal swab collection for laboratory confirmation. The samples were analyzed with the Rapid Antigen Test (RAgT), conducted on-site, using the Celer Wondfo SARS-CoV-2 test¹⁵, or sent to the Respiratory Viruses Lab (IOC/Fiocruz) for Reverse Transcription followed by Polymerase Chain Reaction (RT-PCR).

Eligibility criteria

Patients who were 15 years of age and older, with a suspicion of COVID-19 and a positive result in a confirmatory diagnostic test, collected up to the seventh day of the onset of the symptoms, were considered eligible. Records of patients whose consultations occurred up to 90 days after a positive test for COVID-19 were excluded from the analysis.

Data sources

Information regarding patient consultations was recorded in a spreadsheet created by the Health Surveillance Center (HSC) at the PHC unit. Thus, data, such as sex, age, pre-existing health conditions, symptoms and results of COVID-19 diagnostic tests, were obtained from that database (BDNVS).

Information on the predominant viral variant in the municipality of Rio de Janeiro on the date of the onset of symptoms was added to the database. According to the Genomic Surveillance Report by the Municipal Health Secretary of Rio de Janeiro (MHS-RJ)¹⁶, after the initial predominance of the B.1.1.33 and Zeta strains, the predominant variants were Gamma (February to July 2021), Delta (August to November 2021), and Omicron (after December 2021).

Although information on race/skin color was recorded in the consultation records, this was not entered into the local electronic spreadsheet (BDNVS) but rather recorded by the NVS into the e-SUS Notifica, a system developed by the Ministry of Health to report suspected and confirmed cases of COVID-19. The data on race/ skin color was then recovered by integrating the BDNVS with the e-SUS Notifica, using the notification number as the linking key. This process also enabled the recovery of other important information, such as the individual’s Taxpayer Registration Number (CPF) and name of the patient’s mother, used in the process or pairing with other databases to create the variable “outcome”, which corresponds to the worsening of the patient’s health conditions, as well as to obtain information on COVID-19 vaccine doses received by these patients.

To create the variable “worsening”, information was obtained regarding hospitalization and/or death of patients diagnosed with COVID-19. Reports of SARS (regardless of etiology) and of COVID-19, which required hospitalization, were obtained from the Influenza Epidemiological Surveillance Information System (Sistema de Informações da Vigilância Epidemiológica da Gripe - SIVEP-Gripe). Information on deaths used in this study was obtained from the Mortality Information System (Sistema de Informações sobre Mortalidade - SIM). We selected all deaths that occurred in the state of Rio de Janeiro where COVID-19 was recorded in any field (basic cause of death, lines A, B, C, D, or part 2), identified by the codes B342, U071, and U072.

Finally, data on patients’ vaccination status, an important marker for the prevention of hospitalization and reduction of mortality, was included in our analysis through access to information provided by the National Immunization Program Information System (Sistema de Informações do Programa Nacional de Imunização - SI-PNI).

Implementation of the linkage of databases

After the inclusion of information on race/skin color, CPF, and name of the patient’s mother in the BDNVS, we proceeded with the linkage of the different data sources by using the following identification keys: for the SI-PNI, CPF and full name with the date of birth; for the SIVEP-Gripe, CPF, full name, and date of birth; and for the SIM, full name and date of birth. Record linkage using patients’ names and dates of birth was performed based on seven identification keys (Chart 1), representing different levels of approximation. Information on the name of the patient’s mother was used only to confirm identified data.

Chart 1
Description of identification keys used for data linkage between information systems.

Key	Composition
KEY1	Full name + DB1
KEY2	First name + last name + DB1
KEY3	First name + last name + month of DB1 + year of DB1
KEY4	First 3 letters of first name + first 3 letters of last name +month of DB1 + year of DB1
KEY5	First 3 letters of 1st name + middle name + DB1
KEY6	First 3 letters of 1st name + First 3 letters of last name + day of DB1 + year of DB1
KEY7	First 3 letters of 1st name + first 3 letters of the last name + day of DB1 + month of DB1

• Note: ¹DB = date of birth.
• Source: Authors (2024).

Response variable and explanatory variables

The response variable “worsening” was defined as a composite outcome, consisting of the occurrence of SARS (notification in Sivep-Gripe)¹⁴ within 15 days or death from COVID-19 (registered in SIM) within 30 days after symptom onset¹².

The explanatory variables were selected considering specialized literature and information available in the data sources. Our study included data concerning the patients’ demographic characteristics, self-reported prior health conditions, self-reported symptoms of the episode, and a time-dependent variable indicative of the viral variant in the municipality in the epidemiological week of symptom onset.

Therefore, the individual variables included: sex at birth (female, male); age group (15-29, 30-44, 45-59, ≥60); self-reported diagnosis of diabetes mellitus, arterial hypertension, chronic pulmonary disease (asthma/bronchitis, pulmonary emphysema), chronic heart disease, immunosuppression or pregnancy; and symptoms of sore throat, dyspnea, fever, cough, cephalea, loss of taste, loss of smell, runny nose, arthralgia, myalgia, diarrhea, abdominal pain, vomiting, and fatigue. Still at the individual level, we included a variable indicating obesity, as identified by a physician or nurse, and another related to the vaccination status of the patients who were considered “immunized” after receiving a single dose of the Janssen vaccine or two doses of the AstraZeneca, CoronaVac, or Pfizer/BioNTech vaccines in the period of 15 days to six months prior to care¹⁷. The age variable was categorized in four age groups with intervals of 15 years, covering the population segments “young”, “middle-aged”, and “elderly”, with the purpose of facilitating the interpretation of results, based on the age group most commonly present in studies regarding burden of disease.

The time-dependent variable related to the predominant variant at symptom onset was a categoric variable representing four periods: Pre-Gamma (August 2020 to January 2021), Gamma (February to July 2021), Delta (August to November 2021), and Omicron (from December 2021 on).

The variable referring to race/skin color was not included in the analyses for two reasons. First, data was missing in 17% of the cases, making imputation impossible, given the absence of information about other socioeconomic variables, such as income and education. The second reason was the low reliability of this information, evidenced by a divergence exceeding 29% between the NVS and SI-PNI databases.

Statistical analysis

The association between the explanatory variables and the worsening of the disease was investigated using the logistic regression model with the Firth correction¹⁸. This method, which uses a penalized likelihood function, is recommended to reduce bias in the estimates of the parameters and their respective standard errors due to the reduced number of events, especially in small samples. In the case of logistic regression, the Firth method has the advantage of producing finite and consistent estimates of the model’s parameters in scenarios in which such estimates are not obtained by the maximum likelihood method due to complete or nearly complete separation in some explanatory variables of the model.

All the variables that were statistically significant at the level of 25% (p-value<0.25)¹⁹ in the non-adjusted models were assessed for multicollinearity using the Variance Inflation Factor (VIF). Variables with VIF below 2.5²⁰ were then included in the adjusted model, and those significant at the 5% level were subsequently accessed to determine the presence of interactions. The quality of the fit of the model was verified using the Hosmer and Lemeshow test¹⁹, and its accuracy was estimated by the area under the receiver operating characteristic curve (AUC ROC), and by estimating Sensitivity and Specificity at the optimal threshold and at the cutoff point of 0.5¹⁹. Additional analyses were also conducted, excluding the period of the Omicron variant, due to the absence of cases of worsening during this period.

All data processing was performed in the R program environment, version 4.2.1 (http://vv.r-project.org.br/), using the packages: caret, dplyr, epiR, ggplot2, gtsummary, logistf, lubridate, performance, pROC and ResourceSelection.

This study was approved by the Research Ethics Committee of the Escola Nacional de Saúde Pública Sérgio Arouca, CAAE: 56708422.7.0000.5240.

Results

From August 26th, 2020, to January 31st, 2022, 5,493 diagnostic tests were performed at the health unit on individuals aged 15 years and older, treated for a suspicion of COVID-19. Of that total, 1,708 were confirmed by RT-PCR or RAgT and included in the study cohort (Figure 1). The linkage with the SIVEP-Gripe and SIM databases enabled the identification, of 72 (4.2%) cases of worsening (SARS) and 17 (1.0%) deaths, respectively. All deaths were also recorded in the SIVEP-Gripe, with no out-of-hospital deaths identified.

Figure 1
Flowchart of study participant selection.

Table 1 shows the characteristics of the study population according to the occurrence of worsening/SARS. Most of the consultations for COVID-19 were for females (60.9%), with an average age of 42.9 (±15.7) years. Nearly one-third of the patients (32.7%) had been immunized at the time of consultation, but only 9.7% among the cases that worsened. The proportions of worsening during the Pre-Gamma, Gamma, and Delta periods were, respectively, 3.2%, 10.3%, and 4.4%. Although 35.2% of the COVID-19 episodes occurred during the predominance of the Omicron variant, no cases of worsening were observed during that period.

Table 1
Distribution of patient characteristics and predominant viral variants in confirmed COVID-19 cases treated at the PHC unit, according to the occurrence of disease worsening. CSEGSF. August 26th, 2020, to January 31st, 2022.

Variables		Total N=1,708	Non-worsening N=1,636	Worsening N=72	p-value1
Individual characteristics
Sex, n (%)
	Female	1,041 (60.9)	1.001 (61.2)	40 (55.6)	0.404
	Male	667 (39.1)	635 (38.8)	32 (44.4)
Age, mean (SD)		42.86 (15.69)	42.29 (15.41)	55.78 (16.41)	<0.001
Age group, n (%)					<0.001
	15 to 29 years	409 (23.9)	404 (24.7)	5 (6.9)
	30 to 44 years	576 (33.7)	565 (34.5)	11 (15.3)
	45 to 59 years	431 (25.2)	408 (24.9)	23 (31.9)
	60 years and over	292 (17.1)	259 (15.8)	33 (45.8)
Symptoms
Sore throat, n (%)		743 (43.5)	725 (44.3)	18 (25.0)	0.002
Dyspnea, n (%)		156 (9.1)	124 (7.6)	32 (44.4)	<0.001
Fever, n (%)		896 (52.5)	853 (52.1)	43 (59.7)	0.254
Cough, n (%)		1,249 (73.1)	1,191 (72.8)	58 (80.6)	0.188
Cefalea, n (%)		1,120 (65.6)	1,092 (66.7)	28 (38.9)	<0.001
Loss of taste, n (%)		488 (28.6)	475 (29.0)	13 (18.1)	0.059
Loss of smell, n (%)		431 (25.2)	419 (25.6)	12 (16.7)	0.116
Runny nose, n (%)		812 (47.5)	803 (49.1)	9 (12.5)	<0.001
Arthralgia, n (%)		132 (7.7)	130 (7.9)	2 (2.8)	0.167
Myalgia, n (%)		947 (55.4)	915 (55.9)	32 (44.4)	0.072
Diarrheia, n (%)		251 (14.7)	237 (14.5)	14 (19.4)	0.321
Abdominal pain, n (%)		112 (6.6)	105 (6.4)	7 (9.7)	0.387
Vomiting, n (%)		74 (4.3)	72 (4.4)	2 (2.8)	0.714
Fatigue, n (%)		526 (30.8)	487 (29.8)	39 (54.2)	<0.001
Preconditions
Diabetes Mellitus, n (%)		106 (6.2)	89 (5.4)	17 (23.6)	<0.001
Systemic Arterial Hypertension, n (%)		387 (22.7)	357 (21.8)	30 (41.7)	<0.001
Immunosuppression, n (%)		25 (1.5)	24 (1.5)	1 (1.4)	1
Pregnancy, n (%)		15 (0.9)	13 (0.8)	2 (2.8)	0.263
Chronic Heart Disease, n (%)		16 (0.9)	14 (0.9)	2 (2.8)	0.302
Obesity, n (%)		40 (2.3)	25 (1.5)	15 (20.8)	<0.001
Chronic Lung Disease, n (%)		38 (2.2)	36 (2.2)	2 (2.8)	1
Immunized, n (%)		558 (32.7)	551 (33.7)	7 (9.7)	<0.001
Epidemiological indicator
Variant, n (%)					<0.001
	Pre-Gamma	426 (24.9)	412 (25.2)	14 (19.4)
	Gamma	476 (27.9)	427 (26.1)	49 (68.1)
	Delta	205 (12.0)	196 (12.0)	9 (12.5)
	Omicron	601 (35.2)	601 (36.7)	0 (0.0)

• ¹Chi-square test, Fisher’s exact test; SD = standard deviation.
• Source: Authors (2024).

The proportion of immunized individuals and the occurrence of severe cases varied substantially throughout the studied period. Figure 2 illustrates such dynamics, showing the number of confirmed cases of COVID-19 (bars), the proportion of severe cases (line), and the proportion of immunized individuals (color of the bars) by month of the study. It is notable that the proportion of severe cases was higher between February and June 2021. From August 2021 on, there was a progressive decrease in the total number of cases as well as in the number of severe cases, which occurred concomitantly with the increase in the proportion of immunized individuals. In January 2022, after the appearance of the Omicron variant, the tendency toward a decrease in incidence was interrupted, with an extraordinary increase in the monthly number of cases. Among the cases cared for in that month, 83.4% had been immunized previously and none showed worsening of the disease.

Figure 2
Monthly distribution of COVID-19 cases treated at the PHC unit, proportion of immunized individuals and of those experiencing disease worsening (August 2020-January 2022).

Table 2 presents the crude and adjusted odds ratio (OR), along with their respective 95% confidence intervals (95%CI) for the set of variables analyzed. Collinearity between the variables included in the adjusted model was not detected. The results of the unadjusted logistic model showed increasing odds of worsening with age group, being significantly higher for the age group of 45 to 59 years and older, as compared to the 15 to 29-year-old age group, which remained after adjusting for other variables. In the adjusted logistic model, four symptoms showed a statistically significant association with the odds of worsening. While dyspnea (OR: 6.98; 95%CI: 3.61-13.58) and fever (OR: 2.39; 95%CI: 1.31-4.44) were indicators of higher risk, runny nose (OR: 0.34; 95%CI: 0.15-0.71) and loss of taste (OR: 0.35; 95%CI: 0.13-0.91) indicated protection. Comparing the results of the non-adjusted models, it was observed that headache and fatigue lost significance after adjustment with other variables, while fever became a statistically significant predictor.

Table 2
Odds ratio (OR) and 95% confidence interval (95%CI) from non-adjusted and adjusted logistic regression models for COVID-19 worsening among confirmed cases. CSEGSF, August 2020 to January 2022.

Variáveis		Non-adjusted model			Adjusted model
		OR	CI95%	p-value	adjusted OR	CI95%	p-value
Sex (ref. female)
	Male	1.26	0.78-2.02	0.332	-	-	-
Age group (ref. 12 to 19 years)
	30 to 44 years	1.50	0.56-4.53	0.432	1.50	0.44-5.88	0.525
	45 to 59 years	4.23	1.78-12.05	0.001	4.08	1.33-15.38	0.012
	60 years and over	9.49	4.11-26.58	<0.001	7.02	2.14-27.75	0.001
Sore Throat (ref. No)		0.43	0.24-0.72	0.001	0.87	0.44-1.67	0.682
Dyspnea (ref. No)		9.75	5.91-16.00	<0.001	6.98	3.61-13.58	0.000
Fever (ref. No)		1.35	0.84-2.20	0.211	2.39	1.31-4.44	0.004
Cough (ref. No)		1.51	0.86-2.81	0.153	1.28	0.65-2.68	0.484
Headache (ref. No)		0.32	0.20-0.51	<0.001	0.68	0.37-1.26	0.224
Loss of taste (ref. No)		0.55	0.29-0.98	0.042	0.35	0.13-0.91	0.031
Loss of smell (ref. No)		0.60	0.31-1.08	0.089	0.91	0.33-2.35	0.856
Runny nose (ref. No)		0.16	0.07-0.29	<0.001	0.34	0.15-0.71	0.003
Arthralgia (ref. No)		0.41	0.08-1.20	0.115	0.92	0.14-3.87	0.922
Mialgia (ref. No)		0.63	0.39-1.01	0.056	0.66	0.35-1.21	0.176
Diarrhea (ref. No)		1.46	0.78-2.57	0.225	0.97	0.44-2.02	0.946
Abdominal pain (ref. No)		1.66	0.70-3.42	0.229	0.96	0.30-2.69	0.937
Vomiting (ref. No)		0.77	0.16-2.28	0.671	-	-	-
Fatigue (ref. No)		2.78	1.74-4.48	<0.001	1.52	0.82-2.81	0.187
Diabetes Mellitus (ref. No)		5.45	2.99-9.55	<0.001	2.77	1.26-5.92	0.012
Systemic Arterial Hypertension (ref. No)		2.57	1.58-4.13	<0.001	1.17	0.60-2.26	0.642
Immunosuppression (ref. No)		1.38	0.15-5.43	0.717	-	-	-
Pregnancy (ref. No)		4.26	0.83-33.49	<0.001	16.79	1.84-111.20	0.015
Chronic Heart Disease (ref. No)		3.97	0.77-13.28	0.090	6.60	1.10-28.38	0.040
Obesity (ref. No)		17.04	8.45-33.49	<0.001	19.71	7.47-52.78	0.000
Chronic lung disease (ref. No)		1.55	0.31-4.77	0.532	-	-	-
Immunized (ref. No)		0.23	0.10-0.45	<0.001	1.20	0.36-3.59	0.755
Variant (ref. Pre-Gamma)
	Gamma	3.29	1.85-6.22	<0.001	2.83	1.40-6.04	0.003
	Delta	1.38	0.58-3.13	0.460	1.56	0.51-4.61	0.430
	Omicron	0.02	0.00-0.17	<0.001	0.02	0.00-0.25	0.001

• Source: Authors (2024).

Regarding previous health conditions, obesity (OR: 19.71; 95%CI: 7.47-52.78), pregnancy (OR: 16.79; 95%CI: 1.84-111.20), chronic heart disease (OR: 6.60; 95%CI: 1.10-28.38), and diabetes (OR: 2.77; 95%CI: 1.26-5.92) remained significant in the adjusted model, proving to be independent predictors of the risk of worsening. Systemic arterial hypertension, although associated with the worsening of the disease in the non-adjusted model, was no longer significant after adjustment for other variables.

Regarding the viral variants, a higher chance of worsening was observed among patients cared for during the period in which the Gamma variant was predominant (OR: 2.83; 95%CI: 1.40-6.04), while the lowest chance was observed in the period in which the Omicron variant was predominant (OR: 0.02; 95%CI: 0.00-0.25). Vaccination status, which was statistically significant, with a protective effect against worsening of the disease in the non-adjusted model, lost its statistical significance after adjustment for other variables.

The quality of adjustment of the model, evaluated by the Hosmer and Lemeshow test, presented a p-value of 0.44, suggesting good adequacy. The AUC ROC was 94.5% (95%CI: 92.5-96.6), indicating an excellent discriminative capacity. However, at the ideal Youden threshold (0.34), the model showed a specificity of 98.6% (95%CI: 97.9-99.1), but a sensitivity of only 45.8% (95%CI: 34.0-58.0). By modifying the threshold to increase sensitivity, it is possible to achieve a sensitivity of 93.1% (95%CI: 84.5-97.7) with 80.6% specificity (95%CI: 78.6-82.5), which would result in a positive predictive value (PPV) of 17.4% (95%CI: 13.7-21.6) and a negative predictive value (NPV) of 99.6% (95%CI: 99.1-99.9) in the study population. The comparison of the model’s accuracy (96.4%; 95%CI: 95.3-97.2) with the hit rate when classifying all the individuals as belonging to the most common category of non-worsening showed no superiority of the model (p-value=0.12).

Due to the absence of cases of worsening in the period of Omicron predominance, we conducted an additional analysis, maintaining only the episodes that occurred before that period. The results are presented in Tables 3 and 4. There were differences in the crude ORs related to immunization and to four symptoms: cough and fever were risk factors for worsening; loss of smell was a protective factor; sore throat and immunization showed no association. No substantial difference was observed in the magnitude or statistical significance of the adjusted association measurements obtained by multiple regression models for that specific period and for the entire timeframe (including the Omicron period).

Table 3
Odds Ratio (OR) and 95% confidence intervals (95% CI) from adjusted and non-adjusted logistic regression models for COVID-19 worsening, among confirmed cases, CSEGSF, August 2020 to November 2021 (excluding the period of Omicron predominance).

Variables		Non-adjusted model			Adjusted model
		OR	95%CI	p-value	adjusted OR	IC95%	p-value
Sex (ref. female)
	Male	1.08	0.67-1.75	0.739	1.26	0.78-2.02	0.332
Age group (ref. 12 to 19 years)
	30 to 44 years	1.49	0.55-4.53	0.439	1.47	0.43-5.73	0.544
	45 to 59 years	4.13	1.72-11.80	0.001	4.04	1.33-15.15	0.012
	Over 60 years old	9.13	3.92-25.70	<0.001	7.01	2.15-27.52	0.001
Sore Throat (ref. No)		0.68	0.39-1.16	0.160	0.86	0.44-1.64	0.654
Dyspnea (ref. No)		8.70	5.20-14.49	<0.001	7.03	3.62-13.74	<0.001
Fever (ref. No)		1.73	1.07-2.83	0.024	2.39	1.32-4.44	0.004
Cough (ref. No)		1.83	1.04-3.43	0.034	1.27	0.64-2.65	0.506
Headache (ref. No)		0.36	0.22-0.59	<0.001	0.68	0.37-1.27	0.225
Loss of taste (ref. No)		0.36	0.19-0.64	<0.001	0.34	0.12-0.89	0.028
Loss of smell (ref. No)		0.37	0.19 -0.67	0.001	0.91	0.33-2.34	0.844
Runny nose (ref. No)		0.22	0.10-0.42	<0.001	0.33	0.15-0.69	<0.001
Arthralgia (ref. No)		0.71	0.14-2.13	0.581	-	-	-
Mialgia (ref. No)		0.63	0.39-1.02	0.061	0.65	0.35-1.20	0.168
Diarrhea (ref. No)		1.29	0.69-2.29	0.409	-	-	-
Abdominal pain (ref. No)		1.44	0.60-2.99	0.389	-	-	-
Vomiting (ref. No)		0.94	0.19-2.89	0.931	-	-	-
Fatigue (ref. No)		2.46	1.53-3.99	<0.001	1.48	0.80-2.74	0.207
Diabetes Mellitus (ref. No)		4.67	2.53-8.32	<0.001	2.81	1.27-6.02	0.011
Systemic Arterial Hypertension (ref. No)		2.49	1.52-4.05	<0.001	1.17	0.61-2.25	0.635
Immunosuppression (ref. No)		0.94	0.10-3.74	0.004	1.47	0.14-8.19	0.706
Pregnancy (ref. No)		4.29	0.80-15.89	0.082	17.31	1.90-115.24	0.014
Chronic Heart Disease (ref. No)		3.16	0.61-10.98	0.150	6.68	1.12-28.76	0.039
Obesity (ref. No)		14.05	6.78-28.78	<0.001	20.64	7.75-55.95	<0.001
Chronic lung disease (ref. Não)		1.60	0.32-5.08	0.515	-	-	-
Immunized (ref. Não)		1.91	0.79-4.03	0.138	1.21	0.37-3.61	0.740
Variant (ref. Pre-Gamma)
	Gamma	3.29	1.85-6.22	<0.001	2.82	1.40-5.98	0.003
	Delta	1.38	0.58-3.13	0.460	1.57	0.51-4.63	0.422
	Omicron

• Source: Authors (2024).

Table 4
Assessment of the discriminative quality of adjusted models when excluding and including the period of Omicron predominance.

	Adjusted model Period without Omicron		Adjusted model Complete period
		CI95%		CI95%
AUC (%)	91.4	88.2-94.5	94.54	92.5-96.6
Accuracy (%)	81.9	79.5-84.2	96.37	95.37-97.21

• Source: Authors (2024).

Discussion

In this study, we describe the characteristics of a cohort of confirmed cases of COVID-19 treated in PHC within seven days of symptom onset, identify occurrences of disease worsening (SARS or death), and evaluate the association between worsening and the variables collected during the first PHC visit. We identified worsening in 4.2% (95%CI: 3.3-5.3) of the cases, with a risk that varied from 0% (95%CI: 0.0-0.6), during the predominance of the Omicron variant, to 10.3% (95%CI: 7.7-13.4) during the predominance of the Gamma variant. After adjustments for other variables, being over 45 years of age; having obesity, chronic heart disease or diabetes mellitus; being pregnant; reporting dyspnea or fever, and becoming sick in the period of predominance of the Gamma variant, were factors associated with a higher risk of worsening. By contrast, having a runny nose or loss of taste were identified as protective factors.

Advanced age is well documented in the literature as a risk factor for hospitalization and death by COVID-19^5,21. In our study, the higher odds of worsening became evident starting from the age group of 45-59 years, when compared to those aged 15-29 years. From 60 years old onwards, the odds of worsening was about 7-fold that of the reference age group.

Regarding pre-existing health conditions, obesity was the factor that most increased the risk of disease worsening. Indeed, the medical literature indicates a growing risk associated with a higher body mass index (BMI), starting from the overweight category, particularly among younger individuals^22-24. Our estimates, however, were higher than those reported by studies in other countries^22-24. An overrepresentation of higher degrees of obesity among the individuals classified as obese may have contributed to increasing the risk associated with this condition. Additionally, specific characteristics of our population and the residual confounding related to race/skin color and socioeconomic status, which were not controlled for in the analyses, may also have influenced the results. There is evidence, for example, that the increased risk associated with excess weight is greater among black individuals when compared to white individuals²⁵.

Pregnancy also proved to be an independent risk factor, considerably increasing the odds of the disease worsening. This result is in line with studies showing that pregnant women have a greater probability of being admitted to intensive care units and receiving mechanical ventilation, than do non-pregnant women, especially when older than 35 years of age^26-28.

Chronic heart disease and diabetes have become well established as important risk factors for disease worsening and death by COVID-19^5,29-31. However, arterial hypertension is reported in a less consistent manner. In general, studies do not demonstrate a significant effect of hypertension after adjusting for age and other comorbidities, suggesting that the higher risk among hypertensive patients may be related to other conditions that are more prevalent in this population group^6,32,33. In our study, in a similar way, hypertension lost significance after adjustment for other variables.

Dyspnea was the main predictive symptom for hospitalization or death, corroborating the findings of other studies^7,9,10. We did not find studies in the literature that addressed runny nose or loss of taste as prognostic factors in confirmed cases of COVID-19; however, it is possible that the presence of these symptoms reflects a pattern of involvement more restricted to the upper airways.

Concerning viral variants, Gamma was identified as a significant risk factor for disease worsening with the highest worsening rate (10.2%) occurring during its predominance. In contrast, during the Omicron period, no worsening cases were identified. These results are similar to what has been reported in studies conducted in Brazil, which identified the second wave of COVID-19, when the Gamma variant was predominant, as the most lethal phase, with an average of 21,000 deaths per epidemiological week³⁴.

Vaccination status, which appeared in the non-adjusted model as a protective factor against the worsening of COVID-19, lost relevance after adjusting for other variables, especially the predominant variant in the period. During the predominance of the Omicron variant, 83.4% of the cases attended were immunized, a number close to the 90% verified for the adult population of the municipality of Rio de Janeiro at the same time^34,35. The absence of severe cases during this study period, most likely resulting from the higher proportion of vaccinated people combined with Omicron’s lower capacity to cause severe cases^11,12,36, compromised the investigation of the interaction between the variables “viral variant” and “immunization”. It also hindered the assessment of the role of immunization as a protective factor against disease worsening. The analyses focused exclusively on the pre-Omicron period did not show any effect of vaccination.

The current study stands out for being based on real-world data, where the environment is not controlled by the researcher, which may offer advantages in terms of applicability but also leads to some limitations. The level of detail and accuracy of the information reflects the workflow of a PHC unit, as well as the deficiencies inherent to the information systems.

The low reliability of race/skin color data, with high disagreement between e-SUS Notifica and SI-PNI, made the use of this variable impractical, despite its recognized relevance as a prognostic predictor, as evidenced by the higher COVID-19 mortality rate among the black population in Brazil^37,38. The information on obesity, an important prognostic factor^22-24, was based on identification by the healthcare professional, without the recording of weight and height data, reflecting the registration model adopted by the e-SUS-VE influenza-like syndrome surveillance system. This method of assessment may have led to the predominant identification of the most severe cases, resulting in overestimation of the risk associated with obesity and compromising comparisons with studies based on BMI, as well as the generalization of our results for this condition.

Considering that ethnicity, income, and education have already been identified as factors associated with COVID-19 worsening^6,13,39, the absence of socioeconomic and race/skin color data in this study represents a significant limitation. In addition to making it unfeasible to analyze their association with worsening, this gap also prevented the adjustment of association measures for other variables based on these characteristics, increasing the risk of residual confounding.

The identification of severe cases was carried out by linking clinical visit data with secondary data from the Information Systems of the Ministry of Health. In this procedure, the sensitivity of outcome identification may be compromised by failures in case matching, caused by the absence of more robust identification keys or by problems in the keys based on names or birth dates, which are susceptible to spelling and typing mistakes. To minimize this problem, in addition to CPF and National Health Registration (Cadastro Nacional de Saúde - CNS), we used seven additional identification keys, combining parts of names and birth dates, which we believe provided high sensitivity in the identification of outcomes. Identification of the disease worsening may also have been affected by underreporting of SARS cases in Sivep-Gripe. However, the main deficiencies in this system are related to incompleteness of information regarding etiology and progression of the cases, rather than the initial registration^40,41.

It is worth noting that in another cohort of cases cared for in PHC in Brazil between April and May 2020, a frequency of SARS of 6.8% was observed⁴², a value within the range that we observed during periods dominated by different variants (0 to 10.2%). In the city of Rio de Janeiro, the average ratio between hospitalized and confirmed cases of COVID-19 in the study period was higher: 7.9%. The lower rate of diagnostic confirmation for mild cases, especially during the first year of the pandemic⁴³, combined with the inclusion of emergence and hospital visits, where patients have a more severe profile, explains this difference⁴².

Although many of the identified risk factors are well established, the relative importance of these factors may vary. Considering that this study was conducted using data from one single PHC unit from a low-income area of Rio de Janeiro, caution is required when generalizing the results for populations with different profiles. The relatively small number of outcomes (SARS or death), characteristic of a low-risk population, such that served by PHC unit, was another limitation of this study. This resulted in less precise estimates, with wide confidence intervals, despite the use of a model better suited for this type of problem.

Unlike most prognostic studies on COVID-19, which are based on hospital records from developed countries^10,44, the current study used information collected during the routine care of a Brazilian PHC unit with a simple form adaptable to similar services. This allowed for the evaluation of risk factors related to the first consultation, in the initial phase of the disease. Moreover, the long study period allowed for the investigation of the effects of viral variants and vaccination status on the occurrence of hospitalization and death.

Although the accuracy of the predictive model was high (96.4%), its sensitivity was low (45.8%), showing that accuracy (proportion of correct answers) may overestimate the discriminative capacity of a model when the predicted event is very rare. The change in the cutoff point to achieve a sensitivity of 93% resulted in a drop in specificity to 80.6% and a positive predictive value of only 17.4% in the study population. If the model were applied to identify high risk patients in this population, this would be equivalent, for example, to monitoring six people for each one that would actually worsen (SARS or death), which could be a gain in monitoring efficiency. However, it is likely that changes resulting from immunization and the emergence of new variants, such as the substantial reduction of risk of worsening coinciding with the emergence of the Omicron variant, may compromise the model’s performance for future cases.

Conclusion

Throughout the pandemic, changes in the genetic characteristics of the virus and in the susceptibility of individuals to infection, whether through immunization or to previous infection, have made the development of prognostic prediction models for COVID-19 cases an increasingly challenging task.

Our study corroborates findings regarding a set of risk factors for disease worsening that are more prevalent in the general population, such as older age, pregnancy, obesity, diabetes, and chronic heart disease. However, it lacked sufficient information to investigate the risks associated with socioeconomic factors, and it did not have the statistical power to identify less frequent risk factors, or the level of protection provided by vaccination. Despite these limitations, it produced a useful model to understand the factors involved in COVID-19 worsening throughout the study period in the context of PHC in Rio de Janeiro.

Future changes in the epidemiological scenario may require new strategies for COVID-19surveillance and control. It is essential to maintain genome and epidemiological surveillance, as well as to ensure efficient communication with health professionals involved in direct patient care. This includes providing alerts about new ways of transmission or clinical presentations of the disease, ensuring timely identification and care for the patients.

References

• ¹ Centers for Disease Control and Prevention (CDC). COVID-19 Clinical Presentation [Internet]. 2024 [cited 2024 maio 20]. Available from: https://www.cdc.gov/covid/hcp/clinical-care/covid19-presentation.html
  » https://www.cdc.gov/covid/hcp/clinical-care/covid19-presentation.html
• ² Centers for Disease Control and Prevention (CDC). Interim clinical guidance for management of patients with confirmed coronavirus disease (COVID-19) [Internet]. [cited 2020 abr 24]. Available from: https://www.cdc.gov/coronavirus/2019-ncov/hcp/clinical-guidance-management-patients.htm
  » https://www.cdc.gov/coronavirus/2019-ncov/hcp/clinical-guidance-management-patients.htm
• ³ Wu Z, McGoogan J. Characteristics of and Important Lessons from the Coronavirus Disease 2019 (COVID-19) Outbreak in China. JAMA 2020; 323(13):1239-1242.
• ⁴ Brasil. Ministério da Saúde (MS). Ministério da Saúde - COVID-19 no Brasil [Internet]. 2022 [acessado 2021 fev 22]. Disponível em: https://susanalitico.saude.gov.br/extensions/covid-19_html/covid-19_html.html
  » https://susanalitico.saude.gov.br/extensions/covid-19_html/covid-19_html.html
• ⁵ Centers for Disease Control and Prevention (CDC). COVID-19 People with Certain Medical Conditions and COVID-19 Risk Factors [Internet]. 2024 [cited 2024 jul 23]. Disponível em: https://www.cdc.gov/covid/risk-factors/index.html
  » https://www.cdc.gov/covid/risk-factors/index.html
• ⁶ Hippisley-Cox J, Khunti K, Sheikh A, Nguyen-Van-Tam JS, Coupland CAC. Risk prediction of covid-19 related death or hospital admission in adults testing positive for SARS-CoV-2 infection during the omicron wave in England (QCOVID4): cohort study. BMJ 2023; 381:e072976.
• ⁷ Sudre CH, Lee K, Lochlaimn M, Varsavsky T, Murray B. Symptom clusters in COVID-19: A potential clinical prediction tool from the COVID Symptom Study app. Sci Adv 2021; 7:eabd4177.
• ⁸ Infection Disease Society of America (IDS) [Internet]. 2024 [cited 2024 jun 21]. Disponível em: https://www.idsociety.org/practice-guideline/covid-19-guideline-treatment-and-management/.
  » https://www.idsociety.org/practice-guideline/covid-19-guideline-treatment-and-management
• ⁹ Wynants L, Calster B, Collins GS, Riley RD, Heinze G, Schuit E, Bonten MMJ, Dahly DL, Damen JAA, Debray TPA, Jong VMT, De Vos M, Dhiman P, Haller MC, Harhay MO, Henckaerts L, Heus P, Kammer M, Kreuzberger N, Lohmann A, Luijken K, Ma J, Martin GP, McLernon DJ, Andaur Navarro CL, Reitsma JB, Sergeant JC, Shi C, Skoetz N, Smits LJM, Snell KIE, Sperrin M, Spijker R, Steyerberg EW, Takada T, Tzoulaki I, Kuijk SMJ, van Bussel B, Horst ICC, Royen FS, Verbakel JY, Wallisch C, Wilkinson J, Wolff R, Hooft L, Moons KGM, Smeden M. Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisal (versão 4). BMJ 2022; 369:m1328.
• ¹⁰ Espinosa-Gonzalez A, Prociuk D, Fiorentino F, Ramtale C, Mi E, Mi E, Glampson B, Neves AL, Okusi C, Husain L, Macartney J, Brown M, Browne B, Warren C, Chowla R, Heaversedge J, Greenhalgh T, de Lusignan S, Mayer E, Delaney BC. Remote COVID-19 Assessment in Primary Care (RECAP) risk prediction tool: derivation and real-world validation studies. Lancet Digit Health 2022; 4(9):e646-e656.
• ¹¹ Menni C, Valdes AM, Polidori L, Antonelli M, Penamakuri S, Nogal A, Louca P, May A, Figueiredo JC, Hu C, Molteni E, Canas L, Österdahl MF, Modat M, Sudre CH, Fox B, Hammers A, Wolf J, Capdevila J, Chan AT, David SP, Steves CJ, Ourselin S, Spector TD. Symptom prevalence, duration, and risk of hospital admission in individuals infected with SARS-CoV-2 during periods of omicron and delta variant dominance: a prospective observational study from the ZOE COVID Study. Lancet 2022; 399(10335):1618-1624.
• ¹² Nyberg T, Ferguson NM, Nash SG, Webster HH, Flaxman S, Andrews N, Hinsley W, Bernal JL, Kall M, Bhatt S, Blomquist P, Zaidi A, Volz E, Aziz NA, Harman K, Funk S, Abbott S; COVID-19 Genomics UK (COG-UK) consortium; Hope R, Charlett A, Chand M, Ghani AC, Seaman SR, Dabrera G, De Angelis D, Presanis AM, Thelwall S. Comparative analysis of the risks of hospitalisation and death associated with SARS-CoV-2 omicron (B.1.1.529) and delta (B.1.617.2) variants in England: a cohort study. Lancet 2022; 399(10332):1303-1312.
• ¹³ Antonelli M, Penfold RS, Merino J, Sudre CH, Molteni E, Berry S, Canas LS, Graham MS, Klaser K, Modat M, Murray B, Kerfoot E, Chen L, Deng J, Österdahl MF, Cheetham NJ, Drew DA, Nguyen LH, Pujol JC, Hu C, Selvachandran S, Polidori L, May A, Wolf J, Chan AT, Hammers A, Duncan EL, Spector TD, Ourselin S, Steves CJ. Risk factors and disease profile of post-vaccination SARS-CoV-2 infection in UK users of the COVID Symptom Study app: a prospective, community-based, nested, case-control study. Lancet Infect Dis 2022; 22(1):43-55.
• ¹⁴ Brasil. Ministério da Saúde (MS). Secretaria de Vigilância em Saúde. Guia de vigilância epidemiológica: emergência de saúde pública de importância nacional pela doença pelo coronavírus 2019 - covid-19. Brasília: MS; 2022.
• ¹⁵ Celer Wondfo SARS-CoV-2 Ag Rapid Test [leaflet]. China: Celer Biotecnologia SA; 2021.
• ¹⁶ Centro de Informações Estratégicas em Vigilância em Saúde Rio. Vigilância Genômica COVID-19 - 2020 a 2022 Semana epidemiológica 19. Rio de Janeiro: SMS-RJ; 2022.
• ¹⁷ Brasil. Ministério da Saúde (MS). Secretaria de Vigilância em Saúde. Plano Nacional de Operacionalização da Vacinação contra a COVID-19. Brasília: MS; 2020.
• ¹⁸ Puhr R, Heinze G, Nold M, Lusa L, Geroldinger A. Firth's logistic regression with rare events: accurate effect estimates and predictions? Stat Med 2017; 36(14):2302-2317.
• ¹⁹ Hosmer DW, Lemeshow S, Sturdivant RX. Applied logistic regression. 3ª ed. Hoboken, New Jersey: Wiley; 2013.
• ²⁰ Alisson PD. Logistic Regression Using the SAS System: Theory and Application [Internet]. SAS; 2001 [cited 2024 jun 21]. Available from: https://www.wiley.com/en-us/Logistic+Regression+Using+the+SAS+System%3A+Theory+and+Application-p-9780471221753
  » https://www.wiley.com/en-us/Logistic+Regression+Using+the+SAS+System%3A+Theory+and+Application-p-9780471221753
• ²¹ COVID-19 Forecasting Team. Variation in the COVID-19 infection-fatality ratio by age, time, and geography during the pre-vaccine era: a systematic analysis. Lancet 2022; 399(10334):1469-1488.
• ²² Matu J, Griffiths A, Shannon OM, Jones A, Day R, Radley D, Feeley A, Mabbs L, Blackshaw J, Sattar N, Ells L. The association between excess weight and COVID-19 outcomes: An umbrella review. Obes Rev 2024; 25(10):e13803.
• ²³ Kadowaki T, Matsumoto N, Matsuo R, Mitsuhashi T, Sasaki A, Takao S, Yorifuji T. Obesity, overweight, and severe prognosis in COVID-19 patients in Japan. J Infect Chemother 2023; 29(12):1109-1113.
• ²⁴ Yang Y, Wang L, Liu J, Fu S, Zhou L, Wang Y. Obesity or increased body mass index and the risk of severe outcomes in patients with COVID-19: A protocol for systematic review and meta-analysis. Medicine 2022; 101(1):e28499.
• ²⁵ Gao M, Piernas C, Astbury NM, Hippisley-Cox J, O'Rahilly S, Aveyard P, Jebb SA. Associations between body-mass index and COVID-19 severity in 6·9 million people in England: a prospective, community-based, cohort study. Lancet Diabetes Endocrinol 2021; 9(6):350-359.
• ²⁶ Gonçalves B, Franco R, Rodrigues A. Maternal mortality associated with COVID-19 in Brazil in 2020 and 2021: Comparison with non-pregnant women and men. PLoS One 2021; 16(12):e0261492.
• ²⁷ Allotey J, Fernandez S, Bonet M, Stallings E, Yap M. Clinical manifestations, risk factors, and maternal and perinatal outcomes of coronavirus disease 2019 in pregnancy: living systematic review and meta-analysis. BMJ 2020; 370:m3320.
• ²⁸ Zambrano L, Ellington S, Strid P, Galang R, Oduyebo T. Update: Characteristics of Symptomatic Women of Reproductive Age with Laboratory-Confirmed SARS-CoV-2 Infection by Pregnancy Status. MMWR Morb Mortal Wkly Rep 2020; 69(44):1641-1647.
• ²⁹ Wang S, Zhu R, Zhang C, Guo Y, Lv M, Zhang C, Bian C, Jiang R, Zhou W, Guo L. Effects of the pre-existing coronary heart disease on the prognosis of COVID-19 patients: A systematic review and meta-analysis. PLoS One 2023; 18(10):e0292021.
• ³⁰ Jackson SL, Woodruff RC, Nagavedu K, Fearrington J, Rolka DB, Twentyman E, Carton TW, Puro J, Denson JL, Kappelman MD, Paranjape A, Thacker D, Weiner MG, Goodman AB, Lekiachvili A, Boehmer TK, Block JP; PCORnet Collaboration Authors. Association Between Hypertension and Diabetes Control and COVID-19 Severity: National Patient-Centered Clinical Research Network, United States, March 2020 to February 2022. J Am Heart Assoc 2023; 12(21):e030240.
• ³¹ Hartmann-Boyce J, Rees K, Perring JC, Kerneis SA, Morris EM, Goyder C, Otunla AA, James OA, Syam NR, Seidu S, Khunti K. Risks of and From SARS-CoV-2 Infection and COVID-19 in People with Diabetes: A Systematic Review of Reviews. Diabetes Care 2021; 44(12):2790-2811.
• ³² Williamson EJ, Walker AJ, Bhaskaran K, Bacon S, Bates C, Morton CE, Curtis HJ, Mehrkar A, Evans D, Inglesby P, Cockburn J, McDonald HI, MacKenna B, Tomlinson L, Douglas IJ, Rentsch CT, Mathur R, Wong AYS, Grieve R, Harrison D, Forbes H, Schultze A, Croker R, Parry J, Hester F, Harper S, Perera R, Evans SJW, Smeeth L, Goldacre B. Factors associated with COVID-19-related death using OpenSAFELY. Nature 2020; 584(7821):430-436.
• ³³ Norman DC. Fever in the Elderly. Clin Infect Dis 2000; 31(1):148-151.
• ³⁴ Moura EC, Cortez-Escalante J, Cavalcante FV, Barreto ICDHC, Sanchez MN, Santos LMP. COVID-19: evolução temporal e imunização nas três ondas epidemiológicas, Brasil, 2020-2022. Rev Saude Publica 2022; 56:105.
• ³⁵ Centro de Operações de Emergência. Covid-19 - Boletim Epidemiológico (2020-2022) [Internet]. Rio de Janeiro: SMS-RJ; 2022.
• ³⁶ HealthData.gov. COVID-19 Reported Patient Impact and Hospital Capacity by State Timeseries (RAW) [Internet]. [cited 2024 jun 21]. Available from: https://healthdata.gov/Hospital/COVID-19-Reported-Patient-Impact-and-Hospital-Capa/g62h-syeh/about_data
  » https://healthdata.gov/Hospital/COVID-19-Reported-Patient-Impact-and-Hospital-Capa/g62h-syeh/about_data
• ³⁷ Nascimento JHF, Andrade AB, Gusmão-Cunha A, Cunha AMG. Trends in the morbidity and mortality of coronavirus disease 2019 in different ethnic groups and gender in a large Brazilian city. J Med Virol 2023; 95(5):e28794.
• ³⁸ Marinho MF, Torrens A, Teixeira R, Brant LCC, Malta DC, Nascimento BR, Ribeiro ALP, Delaney R, de Paula PDCB, Setel P, Sampaio JM, Nogales-Vasconcelos AM. Racial disparity in excess mortality in Brazil during COVID-19 times. Eur J Public Health 2022; 32(1):24-26.
• ³⁹ Baggio JAO, Machado MF, Carmo RF, Armstrong AC, Santos ADD, Souza CDF. COVID-19 in Brazil: spatial risk, social vulnerability, human development, clinical manifestations and predictors of mortality - a retrospective study with data from 59 695 individuals. Epidemiol Infect 2021; 149:e100.
• ⁴⁰ Guedes R, Dutra GJ, Machado C, Palma MA. Avaliação dos dados de mortes por COVID-19 nas bases dos cartórios do RC-Arpen, SIVEP-Gripe e SIM no Brasil em 2020. Cad Saude Publica 2023; 39:e00077222.
• ⁴¹ Silva ACCAC, Luiz RR, Zeitoune RCG, Moraes JR, Prata-Barbosa A, Moreira JPL. A gravidade oculta da pandemia de COVID-19 em crianças e adolescentes no Brasil: uma análise territorial da mortalidade hospitalar. Cien Saude Colet 2024; 29:e02662023.
• ⁴² Leal FE, Mendes-Correa MC, Buss LF, Costa SF, Bizario JCS, Souza SRP, Thomaz O, Tozetto-Mendoza TR, Villas-Boas LS, Oliveira-da Silva LC, Grespan RMZ, Capuani L, Buccheri R, Domingues H, Alexander N, Mayaud P, Sabino EC. Clinical features and natural history of the first 2073 suspected COVID-19 cases in the Corona São Caetano primary care programme: a prospective cohort study. BMJ Open 2021; 11(1):e042745.
• ⁴³ Secretaria Municipal de Saúde do Rio de Janeiro. Epi-Rio. Painel Rio COVID-19 [Internet]. [acessado 2024 jun 21]. Disponível em: https://experience.arcgis.com/experience/38efc69787a346959c931568bd9e2cc4
  » https://experience.arcgis.com/experience/38efc69787a346959c931568bd9e2cc4
• ⁴⁴ Wynants L, Calster BV, Collins GS, Riley RD, Heinze G, Schuit E, Bonten MMJ, Dahly DL, Damen JAA, Debray TPA, Jong VMT, De Vos M, Dhiman P, Haller MC, Harhay MO, Henckaerts L, Heus P, Kammer M, Kreuzberger N, Lohmann A, Luijken K, Ma J, Martin GP, McLernon DJ, Andaur Navarro CL, Reitsma JB, Sergeant JC, Shi C, Skoetz N, Smits LJM, Snell KIE, Sperrin M, Spijker R, Steyerberg EW, Takada T, Tzoulaki I, Kuijk SMJ, Bussel B, Horst ICC, Royen FS, Verbakel JY, Wallisch C, Wilkinson J, Wolff R, Hooft L, Moons KGM, Smeden M. Update to living systematic review on prediction models for diagnosis and prognosis of covid-19. BMJ 2021; 372:n236.

• Chief editors:
  Maria Cecília de Souza Minayo, Romeu Gomes, Antônio Augusto Moura da Silva

Publication Dates

• Publication in this collection
  11 Aug 2025
• Date of issue
  July 2025

History

• Received
  05 Nov 2024
• Accepted
  11 Feb 2025
• Published
  13 Feb 2025