Objective:

To analyze the temporal association between COVID-19-related nonpharmaceutical interventions (NPIs) and the incidence of unidentified-pathogen acute hematogenous osteomyelitis (UP-AHO) in children aged 1 to <5 years and to assess the contribution of seasonal viral circulation.

Approach:

• Study Design: A quasi-experimental, population-based interrupted time-series analysis using data from an exhaustive national hospital surveillance system from January 2015 through August 2023.
• Population: Children aged 1 to <5 years hospitalized in mainland France for UP-AHO. Cases associated with Staphylococcus aureus, Streptococcus pyogenes, or Streptococcus pneumoniae were excluded, as were patients with sickle cell disease or prosthesis-related infections.
• Data Collection: Hospital discharge data were obtained from the French Medicalization of Information Systems Programme (PMSI), with diagnoses identified using International Classification of Diseases, Tenth Revision codes. Population-level circulation of 10 seasonal viruses was also assessed using hospitalization data.
• Incidence Measurement: The primary outcome was monthly UP-AHO incidence per 100,000 children aged 1 to <5 years. Secondary analyses examined incidence by age group and estimated the fraction of UP-AHO incidence potentially attributable to individual seasonal viruses.

Key Findings:

• Overall, 5,619 children were hospitalized with UP-AHO during the study period; the median age was 2 years.
• UP-AHO incidence decreased significantly after NPI implementation in March 2020, with an immediate change of −57.6%, and then increased significantly after school reopening in September 2020, with an immediate change of +101.0%.
• These changes were greater among children aged 1 to <3 years.
• Human rhinovirus was estimated to account for 21.3% of UP-AHO incidence, and varicella-zoster virus for 9.8%; no significant associations were found for the other viruses assessed.
• The findings support a temporal association between seasonal virus circulation and UP-AHO in young children, in a setting where Kingella kingae is considered a leading cause of UP-AHO.

Interpretation:

The study found that UP-AHO incidence in children aged 1 to <5 years changed substantially after COVID-19-related NPIs and school reopening, periods that were also marked by major shifts in seasonal viral circulation. The results support a potential temporal relationship between UP-AHO and human rhinovirus and, to a lesser extent, varicella-zoster virus, particularly among younger children.

Limitations:

• Some UP-AHO cases may have been caused by bacteria other than K. kingae because most cases lacked an identified pathogen.
• Reliance on the PMSI medico-administrative database introduced the possibility of diagnostic coding errors, although national quality-control procedures were in place.
• Because PMSI is hospital based, outpatient cases may have been missed.
• Viral circulation was estimated from hospitalization records rather than individual virologic testing in children with UP-AHO, and changes in testing practices or misclassification between enterovirus and rhinovirus could have affected virus-specific estimates.
• Changes in healthcare-seeking behavior or other co-interventions during the COVID-19 pandemic could not be excluded, although the incidence of the control outcome, acute pyelonephritis, remained stable.

Conclusion:

UP-AHO incidence in children aged 1 to <5 years decreased significantly after NPI implementation in March 2020 and increased after schools reopened in September 2020. These changes were temporally associated with shifts in seasonal viral circulation, particularly human rhinovirus and varicella-zoster virus, providing additional support for a relationship between these viruses and UP-AHO, which is caused principally by K. kingae in young children in this setting. The authors note that experimental studies are needed to clarify the molecular interactions between K. kingae and these viruses.