Predictors of death in COVID-19 vaccine breakthrough infections in Brazil.

Dear editor,

Recently, Whitaker and coworkers reported in the Journal evidences of the COVID-19 vaccine effectiveness in most clinical risk groups, with care to highlight the heterogeneity of seropositivity after 1 or 2 doses in individuals with diabetes, chronic heart disease, chronic liver, severe asthma, morbid obesity, and especially immunosuppressed, in which they observed a reduced S-antibody response and vaccine effectiveness. We read with interest the article, especially because we believe in impact of vaccination against COVID-19 in groups with comorbidities.

Through a retrospective, cross-sectional study, based on data from the SIVEP-Gripe Database, the COVID-19 Immunization State Database, and the local medical reporting system, our analysis identified characteristics that may be associated with increased risk of death in vaccinates hospitalized with COVID-19 in a reference health care center. Our outcome of interest was COVID-19-related death in patients with SARS-CoV-2 infection confirmed by RT-PCR with signs/symptoms appearing 15 days or more after vaccine series completion, a period considered reasonable to establish immunity. Vaccine breakthrough infections are defined as the detection of SARS-CoV-2 RNA or antigen in a respiratory specimen collected from a person ≥14 days after they completed all recommended doses. Following these definitions, the patients were divided into two groups upon admission, confirmation of COVID-19 infection via RT-PCR and completed information about COVID-19 vaccination status: i) breakthrough infection and ii) unvaccinated. All variables were subjected to binary logistic regression to define variables that might predict the different clinical outcomes. To select the variables that would comprise the final model, discriminant analysis was performed with p<0.1, estimated by Rao's score test. The variables that obeyed the predefined criteria were subjected to multivariate analysis, with significance defined as p<0.05. All data were tabulated and analyzed with SPSS version 25 software (SPSS, Inc; Chicago, IL, USA).

Between January 5 and September 12, 2021, 2777 (68.5%) were enrolled (Table 1 ). The unvaccinated patients were predominantly male (56.6%) with a mean age of 51.08 (±15.56) years, and 71.5% had one or more comorbidity. These findings are agreed with previous studies that described the clinical profile of patients hospitalized with COVID-19 since the beginning of pandemic, being the disease severity associated with risk factors such as male gender, advanced age, and the presence of comorbidities.3, 4, 5, 6

Table 1

Characteristics of 2777 patients with COVID-19 admitted to hospital between January 5, 2021 and September 12, 2021, according to COVID-19 immunization status.

	Unvaccinated patients			Vaccine Breakthrough Infection*			p-value	O.R. CI 95%	Min	Max
	N responses	N positive or mean	% or s.d.	N responses	N positive or mean	% or s.d.
Sex
Male	2518	1426	56.6%	259	140	54.1%	–
Female	2518	1092	43.4%	259	119	45.9%	0.426	1.1	0.859	1.435
Age	2518	51,08	15.56	259	73,64	12.21	<0.001	–	–	–
> 60 years	2518	727	28.9%	259	230	88.8%	<0.001	19.538	13.153	29.024
Comorbidities	2518	1801	71.5%	259	247	95.4%	<0.001	8.194	4.562	14.721
Recent childbirth	1801	10	0.6%	247	1	0.4%	0.762	0.728	0.093	5.712
Cardiopathy	1801	1089	60.5%	247	214	86.6%	<0.001	4.24	2.904	6.191
Hematological disorder	1801	31	1.7%	247	2	0.8%	0.286	0.466	0.111	1.96
Liver disorder	1801	31	1.7%	247	3	1.2%	0.559	0.702	0.213	2.314
Asthma	1801	74	4.1%	247	10	4.0%	0.964	0.985	0.502	1.932
Diabetes	1801	584	32.4%	247	108	43.7%	<0.001	1.619	1.236	2.121
Neurological disorder	1801	88	4.9%	247	33	13.4%	<0.001	3.124	2.118	4.876
Pneumopathy	1801	84	4.7%	247	22	8.9%	<0.001	3.152	2.057	4.831
Immunocompromised status	1801	76	4.2%	247	21	8.5%	0.001	2.219	1.354	3.639
Kidney disorder	1801	95	5.3%	247	35	14.2%	0.04	1.669	1.02	2.731
Obesity	1801	623	34.6%	247	45	18.2%	<0.001	0.421	0.301	0.59
Symptoms at hospital admission
Fever	2518	1340	53.2%	259	107	41.3%	<0.001	0.619	0.477	0.802
Cough	2518	1716	68.1%	259	166	64.1%	0.183	0.834	0.639	1.090
Sore throat	2518	353	14.0%	259	33	12.7%	0.571	0.896	0.611	1.312
Dyspnea	2518	2299	91.3%	259	230	88.8%	0.179	0.756	0.501	1.114
Respiratory distress	2518	1743	69.2%	259	177	68.3%	0.77	0.960	0.729	1.264
Low oxygen saturation	2518	2394	95.1%	259	252	97.3%	0.108	1.865	0.861	4.037
Diarrhea	2518	315	12,5%	259	27	10.4%	0.331	0.814	0.537	1.233
Vomiting	2518	138	5.5%	258	20	7.8%	0.134	1.449	0.890	2.359
Abdominal pain	2518	80	3.2%	259	7	2.7%	0.676	0.847	0.387	1.853
Fatigue	2518	1052	41.8%	259	140	54.1%	<0.001	1.639	1.268	2.120
Anosmia	2518	149	5.9%	259	5	1.9%	0.008	0.313	0.127	0.770
Ageusia	2518	140	5.6%	258	6	2.3%	0.027	0.404	0.177	0.925
ICU care required	2518	1390	55.2%	259	137	52.9%	0.477	0.911	0.705	1.178
Death	2485	647	26.0%	245	112	45.7%	0.777	–	–	–
Vaccinated	2518	0	0	259	259	100%	NA	–	–	–

In contrast, the vaccinated patients were an average of 73.64 (±12.21) years old, and 95.4% had comorbidities. Being statically higher in this (p<0.001, in both cases) (Table 1). Correlation between complete vaccination, comorbidities, and advanced age was expected according to the definition of priority groups for immunization, which focused on age and comorbidities. This consideration is crucial when analyzing data from vaccinated patients to avoid erroneous associations between vaccination status and hospitalization. The values for Spearman's correlation coefficient confirmed this observation once it was negatively correlated with complete vaccination series when controlling for age and comorbidities (ρ = - 0.005; p = 0.777).

In our data analysis according to vaccine status to define death risk, the univariate analysis identified age > 60 years, presence of comorbidities, cardiopathy, liver disorder, diabetes, neurological disorder, immunocompromised status, pneumopathy, and kidney disease as predictors of death in unvaccinated patients. The discriminant analyses selected six variables (age > 60 years, female sex, liver disorder, kidney disease, obesity, immunocompromised status) that were statistically significant when subjected to multivariate analysis (Table 2 ). Our data depict a second wave of the pandemic in which different variants of the virus were circulating in the country and concerns addressed the impact on younger adults. A change in age profile was observed particularly after the emergence of the P1 (Gamma) variant in Manaus, with higher mortality seen among hospitalized patients in the 20–59 age range compared to the first wave of COVID-19. Our data correspond precisely to the circulation of the Gamma variant in the city and region and consequent overload of the local health system, as already reported.

Table 2

Predictor variables of death among the 1838 unvaccinated patients with COVID-19 admitted to hospital.

	Cure		Death
	N responses	N positive or mean	% or s.d.	N responses	N positive or mean	% or s.d.	p-value	O.R. CI 95%	Min	Max
UNVACCINATED PATIENT
1. Univariate analysis
Sex
Male	1838	1018	55.4%	647	389	60.1%	1.000	–	–	–
Female	1838	820	44.6%	647	258	39.9%	0.037	0.823	0.686	0.988
Age	1838	48.45	14.93	647	58.9	14.91	<0.001	1.049	1.042	1.056
> 60 years	1838	403	21.9%	647	324	50.1%	<0.001	3.572	2.955	4.318
Comorbidities	1838	1193	64.9%	647	587	90.7%	<0.001	5.289	3.988	7.015
Recent childbirth	1193	8	0.7%	587	2	0.3%	0.39	0.506	0.107	2.392
Cardiopathy	1193	686	57.5%	587	394	67.1%	<0.001	1.509	1.227	1.856
Hematological disorder	1193	21	1.8%	587	9	1.5%	0.727	0.869	0.396	1.909
Liver disorder	1193	13	1.1%	587	18	3.1%	0.004	2.871	1.397	5.901
Asthma	1193	51	4.3%	587	23	3.9%	0.723	0.913	0.552	1.509
Diabetes	1193	361	30.3%	587	218	37.1%	0.004	1.362	1.106	1.677
Neurological disorder	1193	45	3.8%	587	43	7.3%	0.001	2.017	1.311	3.101
Pneumopathy	1193	40	3.4%	587	42	7.2%	<0.001	2.221	1.424	3.466
Immunocompromised status	1193	32	2.7%	587	44	7.5%	<0.001	2.940	1.844	4.688
Kidney disorder	1193	44	3.7%	587	51	8.7%	<0.001	2.485	1.639	3.767
Obesity	1193	412	34.5%	587	199	33.9%	0.791	0.972	0.789	1.198
2. Multivariate analysis
Age	–	–	–	–	–	–	<0.001	1.048	1.04	1.057
Sex (female)	–	–	–	–	–	–	<0.001	0.639	0.516	0.791
Liver disorder	–	–	–	–	–	–	0.001	3.409	1.605	7.24
Kidney disorder	–	–	–	–	–	–	0.001	2.166	1.381	3.396
Obesity	–	–	–	–	–	–	<0.001	1.845	1.447	2.353
Immunocompromised status	–	–	–	–	–	–	<0.001	3.232	1.941	5.381
VACCINETED PATIENTS
1. Univariate analysis
Sex
Male	133	69	51.9%	112	63	56.3%	1	–	–	–
Female	133	64	48.1%	112	49	43.8%	0.494	0.838	0.506	1.390
Age	133	70.26	14.10	112	77.7	8.03	< 0.001	1.064	1.034	1.194
> 60 years	133	108	81.2%	112	110	98.2%	0.001	12.731	2.943	55.071
Comorbidities	133	123	92.5%	112	110	98.2%	0.057	4.472	0.959	20.854
Recent childbirth	123	1	0.8%	110	0	0.0%	1.000	–	–	–
Cardiopathy	123	103	83.7%	110	100	90.9%	0.107	1.942	0.866	4.354
Hematological disorder	123	1	0.8%	110	1	0.9%	0.937	1.119	0.069	18.111
Liver disorder	123	1	0.8%	110	2	1.8%	0.508	2.259	0.202	25.265
Asthma	123	4	3.3%	110	6	5.5%	0.413	1.716	0.471	6.249
Diabetes	123	45	36.6%	110	53	48.2%	0.074	1.612	0.954	2.722
Neurological disorder	123	16	13.0%	110	14	12.7%	0.949	0.975	0.452	2.103
Pneumopathy	123	9	7.3%	110	11	10.0%	0.467	1.407	0.560	3.536
Immunocompromised status	123	10	8.1%	110	11	10.0%	0.619	1.256	0.512	3.082
Kidney disorder	123	10	8.1%	110	22	20.0%	0.011	2.825	1.272	6.273
Obesity	123	22	17.9%	110	21	19.1%	0.813	1.083	0.559	2.101
Time between final vaccine dose and symptom onset	133	86.37	40.45	112	79.08	37.08	0.146	0.995	0.989	1.002
2. Multivariate Analysis
Age	–	–	–	–	–	–	<0.001	1.061	1.029	1.093
Kidney disease	–	–	–	–	–	–	0.011	2.887	1.276	6.529

In contrast, the vaccinated patients in our study presented a different profile. While several comorbidities were important predictors of death in unvaccinated patients, in the vaccinated group only age >59 and the presence of kidney disorders were predictors of death in univariate analysis and remained in the multivariate analysis (Table 2). The impact of renal disease, in its various stages, on the formulation of effective immunity after immunization against COVID-19 has been investigated. As reviewed by Hou et al., vaccines are important tools in the prevention of critical COVID-19 in such patients, however the advanced stage of kidney disease and the use of immunosuppressive agents may influence the efficacy of immunizing and the formulation of neutralizing antibodies.

We believe our data complement the findings of Whitaker et coworker, as they demonstrate the effectiveness of the vaccine even in special groups, such as immunosuppressed and especially in a regimen with at least two doses, and we demonstrated that vaccination with a complete regimen with at least two doses can result in a change in the clinical profile of patients hospitalized by COVID-19.

Once breakthrough infection is expected, it is essential to understand the profile of patients who require hospitalization even after complete COVID-19 vaccination for more effective management. In our sample, breakthrough infections associated with hospitalization and death were more frequent in older patients (age ≥60 years) and those with kidney disorders. Although this study is retrospective, it is based on a large dataset. Our propose is highlighting to individual characteristics that should be considered when caring for patients, in addition to signs and symptoms. Our findings demonstrate how vaccination and nonpharmaceutical interventions have changed the profile of COVID-19 in our population, particularly in terms of predictors of death in hospitalized patients: many comorbidities associated with greater risk in the general population are no longer considered risk factors in vaccinated people.

Author contributions

Conceptualization: CFE, MLN. Methodology: CFE, MLN. Investigation: CFE, CAB, LS, GRFC, MMM, TILS, GFF, GCDS, FQ, LMM, AFN, MDB. Data Curation: CFE, MLN. Writing – Original Draft: CFE, MLN. Writing – Review and Editing: CFE, GFF, FQ, MLMS, MLN. Acquisition of Funding: MLN. All authors read and approved the final manuscript.

Declaration of Competing Interest

CFE and MLN has received research grants from Instituto Butantan, Janssen Vaccines & Prevention B.V., Medicago R&D Inc, and Pfizer/BioNTech SE.