Introduction

In March 2020 New Zealand responded to the COVID-19 pandemic by initiating one of the strictest lockdowns worldwide [1]. Known as going ‘hard’ and ‘early’, the goal was to eliminate community spread of the virus. Border closures, hygiene advice, physical distancing, and masking wearing were utilised alongside tracking and tracing active cases [2]. Community restrictions did not formally lift for more than two and a half years, in September 2022. [3]

The unique environment created in response to COVID-19 presents an opportunity to study the effects of social isolation on unrelated diseases. A modification in the transmission of traditional respiratory pathogens has already been noted, with reductions in influenza and respiratory syncytial virus infections during pandemic lockdowns [4]. Perhaps more surprisingly, Group A streptococcus (GAS) infections fell in response to lockdown measures but then rose higher than pre-pandemic levels with signs of increased virulence in late 2022 [4]. It is unclear whether this represents an ‘immunity debt’ among post-lockdown patients or whether GAS infections are concomitant with other respiratory viral illnesses [5].

New Zealand is established as having one of the highest rates of childhood bone and joint infection (BJI) in the world, with incidence > 40 cases/100,000/year. [6] Cases of acute haematogenous osteomyelitis (AHO) and septic arthritis (SA) are predominantly mediated by Staphylococcus aureus. Inequitable disease burden is seen among indigenous Māori and Pacific children, who experience rates of disease 2-3x times higher than their peers [7]. Case severity ranges from isolated disease treated with intravenous antibiotics to those with septic shock and disseminated infection requiring multiple operations and admission to intensive care [8].

Reasons for high levels of invasive S. aureus disease in NZ are not fully understood. It is hypothesised that socioeconomic deprivation and community overcrowding play a part [9]. Prior to lockdown measures, social distancing was considered a non-modifiable variable when examining childhood BJI.

Epidemiological analysis of childhood BJI during the period of COVID-19 restrictions will help delineate the effects of social distancing and increased hygiene measures on rates of disease. Compared to other regions where lockdowns ceased earlier or were less austere, the prolonged and stringent measures undertaken by the NZ government may have created more profound epidemiologic change [10].

The object of this study is to examine the frequency and severity of presentations of SA and AHO immediately before, during, and after the COVID-19 restrictions in New Zealand. Subgroup analysis will be conducted by pathogen type, with attention to pre-hospital care and disease outcomes.

Methods

Health and Disability Ethics Committee (HDEC) approval was obtained for this work together with institutional review board approval (reference: 19/NTA/46). This cohort study is a retrospective analysis of all cases of suspected osteomyelitis and septic arthritis managed in the Auckland region between 2018 and 2023. Children from newborn to age 15-years were included.

Comprehensive review of electronic clinical records was conducted. Data was collected on patient demographics, disease type, and microbiology. Acute haematogenous Osteomyelitis (AHO) was defined using magnetic resonance imaging (MRI) or computed tomography (CT) and/or positive intra-operative culture or bone biopsy. We defined septic arthritis (SA) based on intraoperative culture results, culture results from aspirate, or operative appearance consistent with infection. Cases of chronic infection, post-viral or reactive arthritis, post-operative infection, cases associated with significant malignancy, or patients with insufficient clinical data for analysis have been excluded. Children with a primary diagnosis of osteomyelitis with contiguous infection of the adjacent joint or muscle were included.

Microbiological samples were used from either joint aspirate, blood culture, or intra-operative specimen. Standard agar plate culture was used for synovial fluid aspirate, positive result defined by positive gram stain, cell count > 50,000mm3, growth of pathogen on culture, or intra-operative findings consistent with infection. 16sPCR is also utilised in the Auckland region for selected cases.

Cases were considered to represent severe and complicated disease if the child met specific criteria. This criterion includes admission to a paediatric intensive care unit (PICU), multiple surgeries to control infection, recurrent infection and/or re-admission for further debridement, and children with multifocal sepsis.

Pre-hospital information has been collected including primary care attendance before hospital, any initial misdiagnosis, and duration of symptoms before starting treatment. A patient is considered to be misdiagnosed if the first diagnosis was not AHO or SA. For example, a case would be classified as having initial misdiagnosis if a child with extremity pain was discharged with a diagnosis of transient synovitis, re-presents, and is subsequently diagnosed with infection. Length of stay, number of outpatient clinic visits, and rates of treatment failure have been calculated. General practitioner referral records were reviewed to document recent viral illnesses.

Monthly case numbers have been compared in the period before, during, and after COVID-19 restrictions. Case numbers were divided into subgroups by causative pathogen and severity of illness.

The timeframes for analysis are defined by the periods of COVID-19 mandates initiated by the NZ government. ‘Pre-pandemic’ cases analysed in this study are those presenting between January 2018 and March 2020 [3]. A period of social isolation commenced toward the end of March 2020 and general community restrictions were not lifted until September 2022. The strictest lockdowns were enforced between March-May 2020 and August-December 2021. During these strict lockdowns, educational facilities, daycares, and public venues were generally closed and non-urgent elective surgeries were canceled. A list of regulations for early childhood education centres has been provided in Table 1. This is not exhaustive, but provides an example from public health guidelines published in September 2020 [11]. Additional legislation was passed to facilitate enforcement of contact tracing and other public health regulations [12]. Outside of strict lockdowns, there were still restrictions in the community including mask wearing, limits on gatherings, and hand hygiene rules [1]. ‘Post-pandemic’ cases are defined between October 2022 and August 2023.

Table 1 Examples of COVID-19 restrictions in Education and Childcare facilities

Results

An initial 994 cases were identified by clinical coding. Of these, 563 met criteria for acute BJI (Table 1). The majority had a primary diagnosis of AHO (77%). Staphylococcus aureus was the most likely causative pathogen followed by culture negative cases. Kingella kingae was identified in 23 samples using PCR (7.4% of all positive samples).

Complicated and/or severe illness was seen in 58% of cases; this was most likely due to readmission for treatment failure (n = 106, 19%), multifocal sepsis (n = 90, 16%), and children requiring multiple operations to control disease (n = 85, 15%). Of note, 39 children (7%) were managed in intensive care. A total of 30 children (5%) developed chronic infection.

Table 2 Cohort characteristics

Between January 2018 and March 2020, prior to the COVID-19 pandemic, there were an average 10.1 cases/month (+/-4). During the period of COVID-19 restrictions, ending September 2022, case numbers dropped significantly to 7.9/month (+/-2.5) (p = 0.006) (Table 3; Fig. 1). Proportion of children admitted aged between 6 and 48 months appeared to reduce but did not reach statistical significance (27% vs. 34%, p = 0.16).

Fig. 1
figure 1

Total number of cases by severity of restrictions, 2018–2023. Green: No community restrictions (NB: mask wearing in healthcare settings such as hospitals remained in place until August 2023). Red/Orange-Periods of highly restricted social interaction. Significant dates as follows: 1) March 20th -June 8th 2020 alert levels between 2-4 Nationally. 2) August 12th - October 7th 2020 alert levels between 2-3 in Auckland region. 3) February 14th -March 12th 2021 alert levels between 2-3 in the Auckland region. 4) August 17th – Dec 2nd 2021 alert levels between 2-4 in Auckland region after which alert system abolished, traffic light setting commenced. 5) Dec 15th 2021 regional Auckland border lifted. 6) Sept 12th 2022 community restrictions ended. Yellow: Periods of mildly restricted social interaction, mask wearing, vaccine mandates and hand hygiene. Black line: Mean number of cases per month before, during, and after restrictions. Blue line: Total number of cases per month

Table 3 Pre-hospital care, morbidity, and outcomes

Cases with microbiologic samples positive for S. aureus or S. pyogenes did not reduce significantly during this time (4.2 vs. 3.8, p = 0.4). Conversely, rates of culture negative BJI and cases of K. kingae showed statistically significant reduction (4.2 vs. 2.9, p = 0.04) (Fig. 2). These were combined in analysis as K. kingae is the most common cause of culture-negative septic arthritis.

Fig. 2
figure 2

Monthly Case Numbers Separated by Causative Organism, 2018–2023

With respect to disease severity, there were 5.6 cases/ month (+/- 2.6) classed as severe or complicated prior to COVID-19 restrictions. Frequency of severe illness underwent a statistically significant reduction during the COVID-19 period, dropping to 4.1 cases/month (+/- 1.5, p = 0.01) (Fig. 3).

Fig. 3
figure 3

Monthly Cases of Bone and Joint Infection by Disease Severity, 2018–2023

Examining pre-hospital care, a greater proportion of families presented to primary care during the pandemic period compared to the year prior (45% vs. 64%, p = < 0.05) and the duration of symptoms prior to treatment initiation was longer (5.9 vs. 4.9 days, p = 0.03). The rate of misdiagnosis was unchanged (38% vs. 39%, p = 0.7). Considering the entire cohort, almost half of children reported a recent viral illness before being diagnosed with BJI (46%).

Median length of stay and number of outpatient clinics were not statistically different. However, the number of children who experienced recurrent or chronic infection reduced significantly from pre-pandemic period (9% vs. 4%, p = 0.03).

Discussion

Restrictions introduced during the COVID-19 pandemic significantly reduced rates of childhood osteomyelitis and septic arthritis in the Auckland region. The effects of this continue to be evident as the average number of cases per month has not returned to pre-pandemic levels.

Severity of disease was also lowered during pandemic restrictions. There was a statistically significant reduction in the monthly number of children with multifocal sepsis, PICU admissions, and those requiring multiple operations. Long-term outcomes potentially improved, with recurrence rate reduced by more than half during the COVID-19 period.

Culture negative cases were disproportionately affected compared to cases secondary to S. aureus and S. pyogenes. Whilst both appeared to fall during the pandemic, only culture negative cases or those mediated by K. kingae achieved statistically significant reduction.

Primary healthcare utilisation increased during the COVID-19 period. This may have contributed to the longer average duration of symptoms prior to treatment initiation, which rose during the pandemic period by approximately one day.

There are several possible reasons for the reduced rate and severity of disease noted by our research. Firstly, improved hygiene and mask wearing may have contributed to the reduction in culture negative cases. Recent studies have hypothesised that the majority of culture negative BJI may be attributable to K. kingae, particularly in younger children. This organism is transmitted via respiratory secretions and saliva. Like other respiratory pathogens, the transmission is effectively lowered by mask wearing. [13, 14] In settings where K. kingae has colonised the oropharynx, viral infections of the upper respiratory tract damage the mucosal layer and facilitate K. kingae penetration into the bloodstream and skeletal system. Reduced incidence of respiratory viruses may therefore reduce burden of disease secondary to this pathogen. [15]

Secondly, a reduction in culture-positive disease secondary to S. aureus or S. pyogenes may be in part a direct consequence of reduced skin contact in this period of social isolation. Family gatherings, reduced school and day care attendance, and public social distancing all occurred during the study period. [16]

However, another proposed aetiology of infection suggests that GAS or Staphylococcal infections can be concomitant with respiratory viruses [4, 5]. Importantly, this means that a viral infection may play a role in converting colonisation to invasive disease. It should be reiterated that 46% of children in this study reported a viral infection in the weeks leading up to their admission with invasive BJI. Respiratory virus transmission, particularly influenza and rhinovirus, reduced significantly in NZ during the pandemic [17]. Future studies are required to better characterise the relationship between respiratory pathogens and S. aureus mediated osteomyelitis and septic arthritis.

The two major strengths of this research include unique study setting and thorough evaluation of individual-level data. Compared to other high-income countries, NZ experienced negative excess mortality. The elimination strategy was austere, effective, and relied on radical changes in social structure. For the purposes of this research, it can be determined that limiting gatherings in schools, day care facilities, and public venues lowered childhood BJI admissions from 10.1 to 7.9 cases/month [1, 2, 10].

Secondly, the evaluation of individual level data has not yet been performed for childhood BJI in the setting of COVID-19. Other reports of reduced infection rates have been limited by simple reporting of diagnosis related group information. This does not allow for an assessment of pre-hospital care, microbiology, or treatment outcomes [18,19,20,21]. In particular, recent viral infections were more reliably documented in general practitioner records.

This study is limited by the quality of recorded electronic case information. It should be noted that 16s PCR, although available in our region, is not routinely performed in every case, leading to potential underreporting of K. kingae. [15, 22]

In conclusion, the COVID-19 pandemic in New Zealand provides a unique opportunity to evaluate risk factors for childhood bone and joint disease. Measures undertaken to eliminate spread of COVID-19 coincided with a significant reduction in frequency and severity of presentations for childhood BJI.