Antibodies against SARS-CoV-2 in children and adolescents: a one-year longitudinal study

Abstract

Background

The humoral response to SARS-CoV-2 is not fully understood, especially in children. This study provides valuable insights into the durability of adaptive immunity in paediatric patients who were unvaccinated and naturally infected with the prototype strain of SARS-CoV-2, including those who were asymptomatic or had pre-existing chronic and rare disorders. This study aimed to analyse the antibody response to SARS-CoV-2 during infection and over a 1-year follow-up period in patients aged 0–18 years who were admitted to a tertiary paediatric hospital at the onset of the COVID-19 pandemic.

Methods

This is a descriptive, prospective 1-year cohort study carried out in children and adolescents hospitalised for different reasons from July to October 2020 who presented with SARS-CoV-2 infection. Clinical and serological data (total antibody pool) were collected during SARS-CoV-2 infection and throughout the subsequent year.

Results

During the study period, 122 patients with confirmed SARS-CoV-2 infection were included. Most patients (85%) had at least one serology assessed and reactive throughout 1-year follow-up, even those with asymptomatic infection or immunosuppressive conditions. The mean antibody titre levels reached their maximum value about 60 days after the initial SARS-CoV-2 infection. By the end of the 1-year follow-up, 72.46% of the 69 assessed individuals still had detectable serum antibodies. Patients with comorbidities and/or immunosuppression conditions had lower median titre levels over the year. Subjects with severe clinical presentation of COVID-19 had higher levels of antibody values at the 1-year time point assessment.

Conclusions

This study demonstrated that most subjects presented antibody response against SARS-CoV-2 over 1 year of follow-up, even if they had asymptomatic infection or comorbidities, including patients with immunosuppression. There was a difference in the magnitude and duration of antibody response, lower in patients with comorbidities and longer and higher in subjects who had severe COVID-19 clinical presentation.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Children infected with SARS-CoV-2 develop antibodies, but there is limited evidence on the durability and variation of these responses, particularly in those with comorbidities or immunosuppression.

WHAT THIS STUDY ADDS

• This study shows that most children infected by SARS-CoV-2 maintain detectable antibodies over 1 year, including those with asymptomatic infections or immunosuppression. However, there was a difference in magnitude and duration of antibody response, lower in patients with immunosuppression and longer in severe COVID-19 cases.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The findings provide insights into the long-term humoral immunity of children, highlighting the importance of tailored vaccination strategies and monitoring programmes for paediatric populations, especially those with comorbidities or severe disease.

Introduction

There are knowledge gaps in host immune responses among SARS-CoV-2 recovered patients, especially in children with chronic diseases. The magnitude of the antibody responses can differ and be discordant among individuals and is influenced by disease severity.¹ Monitoring immunity levels is important to guide public health strategies, including vaccination and infection control measures. Furthermore, understanding the persistence of antibodies over time following acute infection in unvaccinated children is crucial for elucidating how humoral immunity can correlate with protection against reinfection and its severity.²

This study aimed to analyse the antibody response to SARS-CoV-2 during infection and over a 1-year follow-up period in patients aged 0–18 years who were admitted to a tertiary paediatric hospital at the onset of the COVID-19 pandemic, comparing the humoral immune response in individuals with and without comorbidities and different clinical presentations.

Methods

Study design and setting

A descriptive, prospective cohort study, conducted over 1 year, examined total antibodies from inpatients aged 0–18 years with confirmed prototype SARS-CoV-2 infection hospitalised at a tertiary paediatric hospital in Brazil. At the time of the study, institutional protocol mandated nasopharyngeal RT-PCR testing for SARS-CoV-2 at admission, regardless of the reason for hospitalisation. The subjects were recruited from July to October 2020 and had blood tests collected at days 01 (Dx), 07, 14, 30, 60, 90, 180 and 365, approximately. As the COVID-19 pandemic began in Brazil in March 2020, the study was planned and received ethics committee approval shortly thereafter, with data collection commencing in July. The decision to collect data over 4 months was informed by the absence of epidemiological data on the incidence and progression of paediatric COVID-19 cases at that time. Furthermore, a convenience sample comprising all eligible paediatric patients treated during this interval was adopted, given the operational and resource constraints characteristic of the early pandemic period.

Patient and public involvement

Patients and members of the public were not involved in the design, conduct and analysis of the study.

Participants and variables

A convenience sample of hospitalised patients aged 0–18 years with active or recent SARS-CoV-2 infection was recruited with consent between July and October 2020. Each patient provided at least one serology sample during or within 1 year following infection. SARS-CoV-2 infection was confirmed by (1) detection of SARS-CoV-2 RNA in a biological sample from the respiratory tract (RT-PCR), independently of clinical symptoms and/or (2) positive serological tests in symptomatic patients, as the study recruitment occurred during the first months of the COVID-19 pandemic.

Clinical data were collected from the electronic medical records, complemented by phone calls over 1 year of follow-up.

Symptomatic patients were categorised into three groups based on their initial clinical presentation: influenza-like syndrome, severe acute respiratory syndrome (SARS) and multisystem inflammatory syndrome in children (MIS-C).

SARS was defined as influenza-like symptoms accompanied by signs of respiratory distress (eg, nasal flaring, intercostal retractions or hypoxaemia).³

MIS-C cases were defined using criteria from the Brazilian Pediatrics Society, which are adapted from the World Health Organization (WHO). MIS-C in individuals aged 0–19 years is defined by:

• Fever ≥3 days.

• Plus at least two of the following:

  • Rash, bilateral non-purulent conjunctivitis or mucocutaneous inflammation (oral, hands or feet).

  • Hypotension or shock.

  • Cardiac involvement (eg, myocarditis, pericarditis, valvulitis, coronary abnormalities, elevated troponin or N-terminal pro-brain natriuretic peptide -NT-proBNP).

  • Coagulopathy (eg, elevated prothrombin time (PT), activated partial thromboplastin time (aPTT), or D-dimer).

  • Acute gastrointestinal symptoms (diarrhoea, vomiting or abdominal pain).

• And elevated inflammatory markers (eg, erythrocyte sedimentation rate (ESR), C-Reactive Protein, or procalcitonin).

• And no other obvious microbial cause, such as bacterial sepsis or toxic shock syndromes.

• And evidence of SARS-CoV-2 infection (positive RT-PCR, antigen or serology) or confirmed contact with a COVID-19 case.⁴

Clinical severity was classified as critical (requiring intensive care unit (ICU) support) or non-critical.

Patients with chronic conditions were defined as long-term health conditions requiring ongoing medical care and continuous follow-up.

Immunocompromised status was assigned to patients with a confirmed diagnosis of primary immunodeficiency or secondary immunosuppression due to treatment (eg, chemotherapy or systemic corticosteroids). The serology was performed in symptomatic or asymptomatic individuals, counted from the date of admission: days 01 (Dx), 07, 14, 30, 60, 90, 180 and 365, within a close interval, depending on the availability of patients studied. The serology tests on day 07 were collected only if the patient was still hospitalised to reduce their displacement to the hospital in a phase of convalescence/contagion. It is worth noting that the collection days are related to the day of hospitalisation (Dx) and not the day of symptom onset.

Diagnostic molecular detection of SARS-CoV-2

Samples from the upper respiratory tract were collected using nasopharyngeal swabs at hospital admission. The molecular detection was performed by the Real-Time Reverse Transcription Polymerase Chain Reaction (rRT-PCR). Viral RNA was extracted using the QIAamp Viral RNA Mini Kit (Qiagen). Briefly, 140 µL of nasopharyngeal swab sample was used, and then RNA was eluted in 60 µL of nuclease-free water. After this step, the reaction and plate configuration were prepared by cycling in the QuantStudio 5 Applied Biosystems with the US-CDC primers and probes.^{5 6}

Serological analysis

Serological tests were performed in blood samples to detect the total pool of antibodies against SARS-CoV-2 by electrochemiluminescence Elecsys Anti-SARS-CoV-2 immunoassay (Roche Diagnostics Int. Ltd).⁷ The assay uses a recombinant protein representing the nucleocapsid (N) antigen in a double-antigen assay format, which favours detection of high-affinity antibodies against SARS-CoV-2 (99.8% specificity). Elecsys Anti-SARS-CoV-2 detects antibody titres, which possibly correlate with neutralising antibodies in neutralisation assays.7,9 It is applicable for qualitative and semi-quantitative detection of a total pool of antibodies (including both IgA and IgG) against SARS-CoV-2 in human serum and plasma. The result is interpreted as positive when the cut-off index (COI) is ≥1.0.

Data handling, analysis and statistics

Quantitative variables were analysed using descriptive measures such as mean, median, SD and IQR. The Research Electronic Data Capture and Microsoft Excel spreadsheets were used for data collection. The IBM SPSS programme, V.23 (2015), was used to perform descriptive and association analyses. The level of significance used in the entire study was 5%. For the association analysis, we calculated Pearson’s χ² test in the exact form, or Monte Carlo simulation, when applicable. To compare the two groups, we used the non-parametric Mann-Whitney U test of independent samples and the Kruskal-Wallis test to three or more groups. The quantitative variables did not present a normal distribution according to the Kolmogorov-Smirnov test.

Results

From July to October 2020, 1132 paediatric patients were admitted to the hospital; 823 subjects had nasopharyngeal swabs rRT-PCR and serology tested, independent of the clinical symptoms or cause of hospitalisation. Among those individuals, 122 patients met the inclusion criteria for the study, representing 10% of the paediatric hospital admissions in the period; 68 (55%) were boys and the mean age was 7.4 years (median 6.0; SD ±5.69 years).

Causes of hospitalisations varied, whether due to underlying disease, chemotherapy, surgeries or severe infectious symptoms. The SARS-CoV-2 infection diagnosis was made by either rRT-PCR detecting the presence of the virus in respiratory swab samples (n=107) or clinical-epidemiological criteria and serological confirmation (n=15). No patients had received COVID-19 vaccine during the recruitment period, as the study period occurred at the onset of the pandemic in Brazil, before the availability of COVID-19 vaccines.

Clinical data

Regarding the syndromic diagnosis, the most common presentation was influenza-like syndrome, present in 43 (35%) individuals, followed by SARS in 25 (20%) cases. Patients with MIS-C represented 17% of the cases (21 cases).

95 (77%) patients had various clinical symptoms, and 27 (22%) were asymptomatic. Table 1 details those symptoms. The most common were fever (68% of the symptomatic patients), vomiting (33%) and cough (32%). 22 symptomatic patients (23%) had signs and symptoms attributed to pre-existing underlying disease at hospital admission rather than SARS-CoV-2 infection.

Table 1. Descriptive analysis of clinical data of children and adolescents with confirmed SARS-CoV-2 virus infection hospitalised in a paediatric hospital in Brazil during the COVID-19 pandemic between July and October 2020.

Variable	No. (%)
Asymptomatic subjects	27 (22.1)
Symptomatic cases	95 (77.9)
Fever measured or reported	65 (68.4)
Vomiting	32 (33.7)
Cough	31 (32.6)
Abdominal pain	25 (26.3)
Dyspnoea	24 (25.3)
Cutaneous manifestations	20 (21.1)
Rhinorrhoeaa	18 (18.9)
Nasal congestion	17 (17.9)
Headache	17 (17.9)
Diarrhoea	16 (16.8)
Myalgia	8 (8.4)
Wheezing	7 (7.4)
Dysgeusia	6 (6.3)
Sneezing	5 (5.3)
Sore throat	5 (5.3)
Anosmia	5 (5.3)
Arthralgia	3 (3.2)
Asthenia	3 (3.2)
Chronic disease	83 (68.0)
Influenza-like syndrome at diagnosis of SARS-CoV-2	43 (35.2)
SARS	25 (20.5)
MIS-C	21 (17.2)
Outcomes
Discharge	118 (96.7)
Death	1 (0.8)
Transfer to another hospital	3 (2.5)
Unit of treatment during the entire current hospital stay
Ward only	82 (67.2)
ICU and ward	35 (28.7)
ICU only	5 (4.1)
Critical COVID-19 cases	30 (30.3)

ICU, intensive care unit; MIS-C, multisystem inflammatory syndrome related to COVID-19 in children; SARS, severe acute respiratory syndrome.

Symptoms at admission persisted for 1–3 days (by the time of admission) in 45% of the subjects, 4–7 days in 38% and ≥14 days in 5% of the children.

This study also analysed patients’ underlying risk factors for viral infections. Most patients had comorbidities (68%, n=83). 66 (54%) were on continuous use of medications at the time of the hospital admission. Half of the patients with chronic conditions (48%, n=40) had immunosuppressive conditions, either due to the disease or the treatment.

During the first 7 days of hospitalisation due to infection, 65.6% of patients (n=80) received some form of pharmacological treatment. The three most frequently administered medications were intravenous antibiotics (n=47; 38.5%), systemic corticosteroids (n=44; 36.1%) and intravenous immunoglobulin (n=23; 18.9%), considering that our hospital was a referral centre for MIS-C at that time.

Regarding respiratory support during the same period, 31 patients (25.4%) required some form of ventilatory assistance. The most used modality was a nasal oxygen cannula, administered to 18.0% (n=22) of patients. 12 patients (9.8%) required invasive mechanical ventilation.

Considering the treatment course, 30% (n=37) of the children needed ICU support because of COVID-19 extreme severity or complications (‘critical cases’). Regarding the outcomes, 117 patients were discharged, 1 patient died due to SARS and 3 were counter-referred to other healthcare facilities (table 1).

Serological data

Most patients (n=104; 85%) had at least one positive serology, demonstrating a serological response to the infection, even those in immunosuppressive conditions. Only five patients did not undergo seroconversion in at least four blood samples over 1 year of follow-up. They had immunosuppression due to severe conditions, detailed in table 2.

Table 2. Descriptive details of five cases that did not have a humoral response over 1 year of follow-up.

Case		Risk factors
	8 years old Collected: Dx. D30, D90, D365	Pneumopathy Hypogammaglobulinaemia Chemotherapy in course as treatment for cancer Use of rituximab
	13 years old Collected: Dx. D14, D30, D90, D180	History of Kawasaki disease and Drug Reaction with Eosinophilia and Systemic Symptoms (DRESS) Undergoing treatment for epilepsy Continuous oral corticosteroid use Use of immunoglobulin
	5 years old Collected: Dx. D30, D90, D365	Chemotherapy in course as treatment for cancer
	15 years old Collected: Dx. D30, D90, D365	Chemotherapy in course as treatment for cancer
	4 months old Collected: Dx. D60, D180, D365	Extreme prematurity    Bronchopulmonary dysplasia    Subglottic stenosis

D7, 7 days after admission; D14, 14 days after admission; D30, approximately 30 days after admission; D60, approximately 60 days after admission; D90, approximately 90 days after admission; D180, approximately 180 days after admission; D365, approximately 1 year after admission; Dx, hospital admission.

Table 3 illustrates the pattern of the humoral response regarding the total antibody titre levels against SARS-CoV-2 and antibody test positivity of the 122 subjects evaluated in eight different moments of infection over 1 year. It was not possible to collect blood samples from all participants at each scheduled time due to challenges in attending the hospital, considering the mobility restrictions imposed during the COVID-19 pandemic.

Table 3. Descriptive analysis of serological data of children and adolescents with confirmed SARS-CoV-2 virus infection, hospitalised in a paediatric hospital in Brazil during the first year of the COVID-19 pandemic.

Timepoint	No. tests performed	No. of tests with antibody detection (%)	Antibody titre levels*
			Mean	Median	SD	Min	Max	IQR
Dx	57	34 (59.64)	41.87	10.36	53.65	0.06	195.40	79.75
D7	25	19 (76.00)	33.11	16.55	38.99	0.06	127.60	57.54
D14	51	40 (78.43)	41.55	18.01	51.47	0.07	204.20	78.84
D30	63	53 (84.12)	40.38	11.77	48.27	0.07	166.10	69.73
D60	52	48 (92.30)	73.25	56.28	67.67	0.06	221.70	127.34
D90	46	41 (89.13)	60.74	24.02	69.18	0.07	225.40	102.70
D180	57	48 (84.21)	54.12	26.85	65.02	0.07	268.00	98.30
D365	69	50 (72.46)	18.78	7.95	33.42	0.07	234.70	22.95

^*Antibody titre levels COI (cut-off index) ≥1.0.

D7, 7 days after admission; D14, 14 days after admission; D30, approximately 30 days after admission; D60, approximately 60 days after admission; D90, approximately 90 days after admission; D180, approximately 180 days after admission; D365, approximately 1 year after admission; Dx, hospital admission.

More than 72% of individuals showed detectable serum antibodies over 1 year, with the highest detection rate observed on D60 at 92.3%, coinciding with the peak average antibody titre of 73.25 (COI ≥1.0). After 1 year of infection, 72.46% of patients still had detectable antibodies, with an average titre of 18.78.

In our cohort, a 7-month-old child died of SARS in the ICU in the first 2 weeks of hospital admission and had no seroconversion in the samples collected. However, no immunological disease was documented by that time.

Comparative data

Group of asymptomatic patients versus symptomatic patients

The median SARS-CoV-2 antibody titre levels of both symptomatic and asymptomatic patients were positive (above the COI), as demonstrated in figure 1. Despite a trend of higher titres in asymptomatic patients at Dx, D14 and D60, this difference was attenuated and inverted 1 year after infection, when the titres were higher in individuals who presented symptoms. The median of D14 serology titre levels was higher in the asymptomatic subjects, and this difference was statistically significant (p=0.009).

Figure 1. Trend line of the mean antibody titre levels of children and adolescents with confirmed SARS-CoV-2 virus infection. Dx, hospital admission; D7, 7 days after admission; D14, 14 days after admission; D30, approximately 30 days after admission; D60, approximately 60 days after admission; D90, approximately 90 days after admission; D180, approximately 180 days after admission; D365, approximately 1 year after admission; SARS, severe acute respiratory syndrome.

Chronic conditions and immunosuppression versus immunocompetent subjects

We found a lower magnitude of antibody response in the group of patients with chronic conditions when compared with immunocompetent individuals (figure 1). In our 1 year cohort, the median antibody titre levels were also lower in the group with comorbidities compared with the others at the time points. This difference was statistically significant at admission (Dx) and over D14, D30 and D365.

When we compared only those in immunosuppressive conditions to the other subjects, the difference between the mean antibody titre levels became even greater (figure 1). The median titre level values were always lower in the immunosuppressed group, and this difference was statistically significant at D30, D60, D90, D180 and D365, counting from the first day of hospital admission.

Regarding seroconversion in these groups, patients with comorbidities or those immunosuppressed were more likely to have non-reactive serology both at admission and 365 days after infection. Patients with immunosuppression were 9.5 times more likely to have a non-reactive D365 serology result (p<0.05, table 4).

Table 4. Association between the D365 serology result and the demographic and clinical data of children and adolescents with confirmed infection by the SARS-CoV-2 virus.

		No. (%) of children and adolescents with confirmed infection			P value	OR	95% CI
		Antibody detected in D365 (n=50)	Antibody not detected in D365 (n=19)	Total (n=69)
Age	<2 years old	10 (20)	1 (5.26)	11 (15.94)	0.166	4.500	0.535 to 37.851
	≥2 years old	40 (80)	18 (94.74)	58 (84.06)
	<10 years old	32 (64)	8 (42.11)	40 (57.97)	0.100	2.444	0.831 to 7.186
	≥10 years	18 (36)	11 (57.89)	29 (42.03)
Chronic disease	Yes	31 (62)	19 (100)	50 (72.46)	0.002	–	–
	No	19 (38)	0 (0)	19 (27.54)
Patient has a chronic disease with immunosuppression either by the disease or by medications	No	19 (57.58)	2 (12.5)	21 (42.86)	0.003	9.500	1.853 to 48.707
	Yes	14 (42.42)	14 (87.5)	28 (57.14)
Influenza-like symptoms at diagnosis of SARS-CoV-2	Yes	21 (42)	5 (26.32)	26 (37.68)	0.230	2.028	0.632 to 6.503
	No	29 (58)	14 (73.68)	43 (62.32)
SARS	Yes	11 (22)	0 (0)	11 (15.94)	0.028	–	–
	No	39 (78)	19 (100)	58 (84.06)
MIS-C	Yes	9 (18)	1 (5.26)	10 (14.49)	0.264	3.951	0.465 to 33.550
	No	41 (82)	18 (94.74)	59 (85.51)
Asymptomatic patient	Yes	7 (14)	5 (26.32)	12 (17.39)	0.289	0.456	0.125 to 1.667
	No	43 (86)	14 (73.68)	57 (82.61)
Outcome	Hospital discharge	49 (98)	19 (100)	68 (98.55)	1.000	–	–
	Transfer to another hospital	1 (2)	0 (0)	1 (1.45)
Place of treatment during the entire current hospital stay	Ward only	33 (66)	17 (89.47)	50 (72.46)	0.081	–	–
	ICU and ward	16 (32)	1 (5.26)	17 (24.64)
	Only ICU	1 (2)	1 (5.26)	2 (2.9)
COVID-19 severity	Critical	17 (34)	1 (5.26)	18 (26.09)	0.029	9.273	1.139 to 75.493
	Non-critical	33 (66)	18 (94.74)	51 (73.91)

p value statistically significant.

D365, approximately 1 year after admission; ICU, intensive care unit; MIS-C, multisystem inflammatory syndrome related to COVID-19 in children; OR, increased odds ratio; SARS, severe acute respiratory syndrome.

Severity of the clinical presentation

The detection of the total pool of antibodies against SARS-CoV-2, considering the severity of the condition of children during hospitalisation, is demonstrated in figure 1. The critically ill children presented maximum mean antibody titre levels later and in greater magnitude when compared with non-critical cases. So, the critical (extremely severe) cases reached the maximum total antibody titres at 6 months (D180) after acute infection, while the non-critically ill ones peaked at D60. Also, those who were critically affected had higher antibody titres at D365. Similar results were found when comparing the humoral response in individuals who presented with SARS; however, in a lower magnitude (figure 1).

In addition, over 1 year, the median antibody titre levels of patients who presented severe COVID-19 symptoms were significantly higher at D180 and D365 than other patients without this clinical presentation.

Therefore, children with severe symptoms of COVID-19 or critical conditions took longer to reach the maximum antibody titre levels. However, they presented high titres later and apparently in a sustained manner.

Similarly, critically ill patients and those with SARS more frequently maintained seroconversion on D365 compared with others (p value <0.05, table 4).

The associative analysis of serology in D365 showed that patients with chronic disease and/or immunocompromised status were more likely to have non-reactive antibody results. On the other hand, patients with SARS and/or critically ill patients were more likely to have a positive D365 serology result. Critically ill COVID-19 patients had a 9.27 times higher chance of a positive result 1 year after the initial hospitalisation (table 4).

Discussion

This study presents a unique cohort of paediatric COVID-19 patients hospitalised in a tertiary centre during the first months of the pandemic, most of whom had pre-existing complex medical conditions and were unvaccinated against SARS-CoV-2. It is significant to point out that, since the study was conducted in a tertiary hospital, many children and adolescents had rare and complex diseases (68% of cases). Furthermore, the participants were hospitalised for various reasons, not exclusively due to infectious symptoms. As a result, 22% of those with confirmed SARS-CoV-2 infection were asymptomatic and were hospitalised due to other conditions, for example, chemotherapy and surgery. The infection detection guided control policies at our institution, a crucial measure during the early stages of the pandemic. We collected blood samples throughout 1 year to perform serological tests measuring total antibody levels, which have been correlated with the levels of neutralising antibodies.^{2 7 10}

Adaptive immunity involves humoral and cellular responses. This study examined the SARS-CoV-2-specific humoral response in hospitalised children and adolescents with different clinical manifestations and severity levels of COVID-19. As many of them had underlying chronic conditions and immunosuppression, a lower seroconversion rate was anticipated. However, most subjects (85%) had antigen-specific antibodies detected during the follow-up, in which 78% of participants had antibodies detected at D14 (from the day of hospital admission) and 72% at the 1-year assessment. It is important to note that we cannot guarantee that the antibodies detected over the 1-year period were solely due to the initial infection, as this is a real-world study conducted during an ongoing pandemic with continuous viral circulation. To minimise this potential bias, periodic telephone follow-ups were conducted with families to identify possible reinfections, and no participants reported a confirmed diagnosis of reinfection.

Our data also demonstrated that the pooled prevalence of positive serology was above 84% within 1 month and over 6 months. There are limited publications on SARS-CoV-2 immunity in healthy paediatric patients.11,13 Tsang et al¹⁴ evaluated adaptive immunity in 31 recovered children and adolescents, focusing on paediatric-specific immune responses to COVID-19. Using two independent assays, they investigated both humoral and cellular immunity. Their findings suggest that most convalescent children and adolescents develop measurable immune responses, with some variability in antibody levels and T-cell activity over time. In that study, SARS-CoV-2-specific receptor-binding domain IgG was detected in recovered patients with a half-life of 121.6 days and an estimated duration of 7.9 months compared with baseline levels in controls. However, that study, in contrast to ours, did not include children with underlying conditions or severe COVID-19.

Our data also demonstrated a sustained humoral response. The mean antibody titre levels peaked approximately 60 days after the initial infection. There has been considerable emphasis on identifying an antibody measurement that is a reliable indicator of protective immunity, known as a correlate of protection (CoP). Sobhani et al^{12 15} evaluated the clinical value of serological testing for SARS-CoV-2 and sought to define immune markers that correlate with protection against the virus. The researchers identified neutralising antibodies as a key CoP, suggesting that measuring these antibodies can help predict immunity and guide public health strategies.¹⁵ It should be emphasised, however, that serological tests for SARS-CoV-2 antibodies do not reflect the richness and durability of immune memory. Dan et al¹⁶ showed that immune memory, including neutralising antibodies and T-cell responses, persisted for at least 8 months in most individuals, with a gradual decline in antibody levels over time. T-cell responses, however, remained more stable, suggesting that while antibody levels may decrease, cellular immunity may provide long-lasting protection against reinfection.

Regarding the variability of the immune response, our findings confirmed that although most immunosuppressed patients demonstrated the presence of antibodies, their response magnitude was lower than others. It is known that immunocompromised children could have a lower immune response than immunocompetent children, both in terms of their capacity to respond to active infections and their response to vaccinations, as exemplified by patients undergoing haematopoietic stem cell transplantation.¹⁷ Furthermore, other studies have shown that immunocompromised patients can mount an immune response, although with lower antibody titres.¹⁸ In our study, all children who failed to seroconvert had immunosuppressive conditions; none were healthy. Martin et al¹⁹ studied 43 subjects, 10 of whom were actively receiving chemotherapy, 10 had previously received chemotherapy and 23 were healthy controls. Paediatric patients receiving chemotherapy and who had previously received chemotherapy had adequate immune activation after both vaccination and acute infection compared with healthy paediatric controls, although there was a demonstrated variability in response for the patients on active chemotherapy. Vaccination against COVID-19 produced superior immune responses compared with acute infection in paediatric patients with cancer and healthy children. Morgans et al²⁰ found that most immunocompromised children did mount a humoral and cellular immune response to the two-dose COVID-19 vaccine series. The majority (86.5%) of the 37 participants had a detectable humoral response after the second and third vaccine doses, with a significant increase in antibody levels after the third dose. Not surprisingly, this response remained lower than what has been demonstrated for healthy children.

We found that antibody detection was more sustained over 1 year of follow-up in patients with severe manifestations during the acute infection compared with those with milder disease, similar to previous published studies.^{12 21} A systematic review by Chivese et al² examined the prevalence of adaptive immunity to COVID-19 and the likelihood of reinfection after recovery, including antibody and T-cell activity, and risk of reinfection over time. Their findings suggest that while adaptive immunity generally protects against reinfection, the level and duration of immunity can vary across individuals and populations. Factors such as age, comorbidities and the emergence of new variants may influence immunity and unfavourable outcomes.

Moreover, we found that children with either symptomatic or asymptomatic infections produced antibodies. These findings were also described in other studies.^{14 22 23} Curiously, the median of D14 serology titre levels was significantly higher in the asymptomatic subjects. This variation is possibly due to the unknown timing of infection onset.

Considering the study limitations, the research design was developed in the early days of the pandemic, with a great lack of information and protocols. As it was a single-centre study, we had limited subjects included. Given the mobility restrictions during the period, many patients faced difficulties collecting follow-up serology samples, which led to missing serology data in some cases. In addition, despite no patient reported confirmed SARS-CoV-2 reinfection during a year of follow-up, which could have modified antibody titres, it could have happened as we did not perform in-person clinical evaluations outside of the periods of scheduled visits and we did not offer molecular diagnostic tests for this purpose. Finally, vaccination was not yet available for the paediatric age group in Brazil for most of the study period. Only four adolescents reported the first dose of COVID-19 shot just before collecting blood at D365. There was no apparent interference from vaccination, as the antibody titres at D365 were lower than those observed in previous collections during the study.

Conclusion

This study demonstrated that most subjects presented antibody response to SARS-CoV-2 over 1 year of follow-up, even if they had asymptomatic infection or severe comorbidities. There was a difference in the magnitude and duration of antibody response, lower in patients with comorbidities and longer in subjects who had severe clinical presentation. Our findings pave the way for large-scale studies, which could help explain the differences in immune response in different groups of COVID-19 patients.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.