📊 epidemiology

Paediatric COVID-19 severity across SARS-CoV-2 variants in hospitalised children in South Africa: a retrospective cohort study

This retrospective cohort study of 354 hospitalized children in South Africa reveals that pediatric COVID-19 severity was highest during the pre-Omicron (Beta and Delta) waves, with younger age, specific variants, and elevated CRP levels identified as key factors associated with severe illness in this resource-limited setting.

Original authors: Fiandrino, S., Di Chiara, C., Dona, D., Dunbar, R., Goussard, P., Lochan, H., Rabie, H., Redfern, A., Truter, C., Van Niekerk, M., van Zyl, G., Verhagen, L. M., van der Zalm, M. M., Paolotti, D.

Published 2026-07-09

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Fiandrino, S., Di Chiara, C., Dona, D., Dunbar, R., Goussard, P., Lochan, H., Rabie, H., Redfern, A., Truter, C., Van Niekerk, M., van Zyl, G., Verhagen, L. M., van der Zalm, M. M., Paolotti, D.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine the SARS-CoV-2 virus as a series of different "characters" in a long-running play, each with its own personality and strength. Some characters are mild, while others are quite aggressive. This study is like a backstage report from a major hospital in Cape Town, South Africa, looking at how these different viral "characters" affected children who ended up in the hospital over a two-year period (from early 2020 to mid-2022).

Here is the story of what the researchers found, broken down simply:

The Setting: A Busy Hospital Stage

The researchers looked at 354 children who were admitted to Tygerberg Hospital. This isn't just any hospital; it's a big referral center that often sees patients with complex health issues, including malnutrition, HIV, and tuberculosis. Because of this, the children in this study represent a group that faces extra challenges, much like a team playing on a difficult, uneven field.

The study divided the timeline into three distinct "seasons" based on which version of the virus was dominant:

The Ancestral Season: The original version of the virus.
The Pre-Omicron Season: A mix of the Beta and Delta versions.
The Omicron Season: The newer, highly contagious version.

The Main Findings: Who Got Sicker and When?

1. The "Pre-Omicron" Season was the Toughest
Think of the Beta and Delta variants as the "heavy hitters" of the virus family. The study found that the hospital was under the most pressure during this season. Children infected during this time were significantly more likely to end up in the Pediatric Intensive Care Unit (PICU) or require serious breathing support compared to the other seasons.

The Analogy: If the virus were a storm, the Pre-Omicron season was the hurricane, while the Omicron season was more like a strong thunderstorm.

2. Age Matters: The Littlest Ones Were Most at Risk
The study confirmed that the youngest children (babies under 1 year old) were the most vulnerable. They were much more likely to need intensive care than older children (ages 1–4).

The Analogy: It's like a house where the foundation is weakest. The infants' bodies were less able to handle the stress of the infection compared to the slightly older kids.

3. The "Omicron" Season was Different
When the Omicron wave hit, the hospital saw a shift. While many children still got sick, the ones who were hospitalized generally didn't get as sick as those in the previous season.

Why? The researchers suggest two reasons: First, the Omicron virus itself might be slightly less aggressive (less "virulent"). Second, by the time Omicron arrived, many people had already been exposed to the virus before, giving their immune systems a bit of a "head start" or shield, even if they weren't vaccinated yet.

The Clues: What Helped Doctors Predict Severity?

The researchers looked for "clues" in the children's blood and symptoms that could predict who would get very sick.

The CRP "Smoke Alarm": They found that a blood test called CRP (C-reactive protein) was a useful clue. CRP is like a smoke alarm in the body; when it goes off high, it means there is a lot of inflammation (fire). Children with very high CRP levels were much more likely to have severe disease.
The Steroid "Bandage": The study noticed that children who received corticosteroids (a type of strong anti-inflammatory medicine) were the ones who got the sickest.
- Important Note: The paper clarifies that this doesn't mean the medicine made them sick. It's the opposite: doctors gave the medicine to the kids who were already in the most trouble. Think of it like seeing a lot of ambulances at a car crash; the ambulances didn't cause the crash, they were there because the crash was bad.

What This Means for the "Play"

This study didn't just count heads; it tried to understand the story of the disease across different times.

The Takeaway: In places with fewer resources (like South Africa), the impact of the virus changes depending on which version is circulating and how old the child is.
The Lesson: To keep children safe, hospitals need to be extra vigilant with babies and watch out for the "heavy hitter" virus versions (like Beta and Delta). They also need to keep an eye on the "smoke alarm" (CRP levels) to spot trouble early.

What the Study Didn't Say

It is important to stick to what the paper actually claims:

It did not say that vaccines were the main reason for the change in severity (since the study period was early in the pandemic for children).
It did not prove that corticosteroids caused better or worse outcomes; it only noted they were used for the sickest kids.
It did not claim these results apply to children who were not in the hospital (since this study only looked at kids who were already admitted).

In short, this paper is a detailed map of how a specific group of children in South Africa navigated three different waves of the pandemic, highlighting that the youngest kids and the "Beta/Delta" waves were the most dangerous combination.

Technical Summary: Paediatric COVID-19 Severity Across SARS-CoV-2 Variants in Hospitalised Children in South Africa

Problem Statement
While children generally experience milder COVID-19 courses than adults, a subset of paediatric patients, particularly infants, develop severe disease requiring intensive care. The epidemiology and clinical severity of SARS-CoV-2 infection in children have evolved with successive Variants of Concern (VOCs). However, existing literature on variant-specific impacts is largely limited to adult populations, mild-to-moderate cases, or high-income healthcare settings. There is a critical knowledge gap regarding how disease severity and clinical presentation have shifted across pandemic waves specifically among hospitalised children in low- and middle-income countries (LMICs), where comorbidities such as malnutrition, HIV, and tuberculosis may exacerbate outcomes.

Methodology
This retrospective cohort study analysed routine care data from the COVID-Kids cohort at Tygerberg Hospital, a secondary and tertiary referral center in Cape Town, South Africa.

Study Population: 354 children (aged 0–15 years) with laboratory-confirmed SARS-CoV-2 infection admitted between March 18, 2020, and June 17, 2022. Infants diagnosed in neonatal services and children with Multisystem Inflammatory Syndrome (MIS-C) without a positive rRT-PCR were excluded.
Variant Stratification: The cohort was divided into three periods based on circulating VOCs:
- Ancestral: March 18 – October 4, 2020.
- Pre-Omicron: November 1, 2020 – October 7, 2021 (combining Beta and Delta variants).
- Omicron: November 17, 2021 – June 17, 2022.
Outcomes:
- Primary Outcome: A composite severe COVID-19 outcome defined as paediatric ICU (PICU) admission, invasive mechanical ventilation, or in-hospital death.
- Secondary Outcome: PICU admission alone.
Statistical Analysis:
- Descriptive statistics compared clinical characteristics across variant periods using $\chi^2$ tests, Fisher–Freeman–Halton exact tests, and Kruskal–Wallis tests.
- Multivariable logistic regression models were employed to assess associations between risk factors and severity. Two models were constructed:
  1. Baseline Model: Adjusted for age, sex, and SARS-CoV-2 variant.
  2. Full Clinical Model: Additionally adjusted for C-reactive protein (CRP) levels, severe dehydration status, and corticosteroid use.
- Missing data were handled via complete-case analysis.

Key Results

Demographics and Comorbidities: The median age was 1.2 years (IQR: 0.3–5.3). No statistically significant differences were observed in sociodemographic characteristics (age, sex) or comorbidity prevalence (HIV, TB, malnutrition) across the three variant periods.
Clinical Presentation: Significant variations were noted in symptoms and signs across variants. The Pre-Omicron period showed higher rates of convulsions, vomiting, tight chest, and severe dehydration compared to other periods. Conversely, the Omicron period was associated with higher rates of non-invasive respiratory support and antibiotic use.
Disease Severity by Variant:
- The Pre-Omicron period exhibited the highest burden of severe disease. Approximately 30% of patients in this period met the composite severe outcome criteria, compared to 17% in the Omicron period and 16% in the Ancestral period.
- PICU admission rates were highest during the Pre-Omicron wave (26.8%), followed by Omicron (16.5%) and Ancestral (14.7%).
- Invasive ventilation and mortality rates were highest in the Pre-Omicron period, though mortality differences were not statistically significant ( $p=0.43$ ).
Risk Factor Analysis (Multivariable Models):
- Age: Infants (<1 year) had significantly higher odds of PICU admission and severe outcomes compared to children aged 1–4 years.
- Variant: Infection during the Pre-Omicron period was independently associated with increased odds of severe disease compared to the Ancestral period (Adjusted Odds Ratio [aOR] = 2.36; 95% CI: 1.21–4.59).
- Clinical Markers: Elevated CRP levels were associated with increased odds of severe disease (aOR = 2.23). Corticosteroid use was strongly associated with severe outcomes (aOR $\approx$ 2.4), which the authors interpret as a marker of disease severity (confounding by indication) rather than a causal treatment effect.

Key Contributions

Longitudinal Characterization: This study provides a systematic characterisation of clinical presentation and risk factors for severe disease across the Ancestral, Pre-Omicron (Beta/Delta), and Omicron waves within a single hospital-based paediatric population in an LMIC setting.
Variant-Specific Severity: It quantifies the shift in severity, identifying the Pre-Omicron period as the peak for severe paediatric outcomes, while noting a reduction in severity during the Omicron wave.
Biomarker Validation: The study reinforces the prognostic value of CRP as an early risk stratification biomarker for severe paediatric COVID-19 in resource-limited settings.
Age-Stratified Risk: It confirms that young age (specifically infancy) remains a critical predictor of severe outcomes across different pandemic waves.

Significance and Claims
The authors claim that these findings contribute to a better understanding of the epidemiology and clinical impact of COVID-19 in children within resource-limited contexts. The study highlights that disease severity is not static but varies significantly by circulating variant and patient age.

Surveillance: The results underscore the necessity for age-stratified, variant-aware surveillance systems in LMICs to monitor shifts in disease burden.
Clinical Triage: The association between young age and severe outcomes supports prioritising infants in clinical monitoring and triage protocols.
Treatment Protocols: The findings regarding CRP and corticosteroid use highlight the need for paediatric-specific trials to establish evidence-based treatment protocols, particularly given the high burden of comorbidities in these settings.
Future Preparedness: The study advocates for the early implementation of standardized data collection protocols to enable robust longitudinal and variant-stratified analyses for future health emergencies.

The authors explicitly note limitations, including the single-center design (potentially limiting generalizability to primary care), the observational nature precluding causal inference, and the inability to include variables such as vaccination status due to data unavailability during the study period.