Article Text
PDF
Abstract
Objective To estimate the contribution of infections to childhood deaths in England and Wales over a 3-year period.
Design Retrospective analysis of national electronic death registration data.
Setting England and Wales.
Patients Children aged 28 days to 15 years who died during 2013–15.
Main outcome measures The proportion of children who died of infection compared with total deaths over 3 years; the main pathogens responsible for infection-related deaths in different age groups; comparison with similar data from 2003 to 2005.
Results There were 5088 death registrations recorded in children aged 28 days to <15 years in England and Wales during the three calendar years, 2013–2015 (17.6 deaths/100 000 children annually) compared with 6897 (23.9/100 000) during 2003–05 (incidence rate ratios (IRR) 0.74, 95% CI 0.71 to 0.77). During 2013–15, there were 951 (18.7%, 951/5088) infection-related deaths compared with 1368 (19.8%, 1368/6897) during 2003–05, equivalent to an infection-related mortality rate of 3.3/100 000 compared with 4.8/100 000 during the two periods (IRR 0.69, 95% CI 0.64 to 0.75), respectively. An underlying comorbidity was recorded in 55.0% (523/951) of death registrations during 2013–15 and increased with age. Where recorded, respiratory tract infection was the most commonly reported presentation (374/876, 42.7%) during 2013–15. Central nervous system infections accounted for only 4.8% (42/876). Overall, 63.1% (378/599) of infection-related deaths were associated with a bacterial, 34.2% (205/599) with a viral and 2.5% (15/599) with a fungal infection.
Conclusions Beyond the neonatal period, all-cause and infection-related childhood mortality rates have declined by 26% and 31%, respectively, over the past decade. However, infection continues to contribute to one in five childhood deaths.
- general paediatrics
- mortality
- vaccine
- infectious diseases
Statistics from Altmetric.com
What is already known on this topic?
The UK has one of the highest childhood mortality rates among industrialised countries.
Infections are associated with significant morbidity and mortality in children.
Vaccination remains one of the most effective strategies for preventing infections.
What this study adds?
Half the infection-related deaths occur in children with underlying comorbidity.
Both the overall and infection-related childhood mortality rates have declined over the past decade.
Infections continue to contribute to one in five childhood deaths in England and Wales.
Introduction
The UK has one of the highest childhood mortality rates across Europe despite a >twofold reduction over the past 25 years, and infection remains a significant cause of death.1 In the UK, we previously reported that infections contributed to 20% of childhood deaths during 2003–05,2 consistent with the Confidential Enquiry into Maternal and Child Health (www.cemach.org) report at the time, which concluded infection to be the ‘largest single cause of death in children dying of an acute physical illness…’. In the 2014 Royal College of Paediatrics and Child Health (RCPCH) report ‘Why children die: death in infants, children and young people in the UK’, infections were estimated to contribute to only 9% of deaths among children aged 1–4 years and 4%–7% in older children,3 most likely because only the primary cause of death was assessed. More importantly, though, the report rarely referred to ‘infection’ as a significant contributor to childhood deaths and, while many of the childhood deaths were considered to be preventable, there was no mention of vaccines in the report.3 Additionally, in a subsequent RCPCH report comparing child health in England with other high-income countries, neither infection prevention nor immunisation was considered among the actions needed to prevent childhood deaths or within the research priorities to improve childhood deaths.4
We used electronic death registration data to estimate the contribution of infections to childhood deaths in England and Wales during 2013–15 and compared our findings with a previous study performed using the same methodology during 2003–05,2 with the aim of identifying key areas of priority for research, interventions and recommendations to help reduce infection-related deaths in children.
Methods
Public Health England receives annual electronic death registration records for England and Wales for surveillance purposes. To estimate the contribution of infections to childhood deaths, anonymised records for children aged 28 days to 15 years between January 2013 and December 2015 were analysed in 2018, when complete and updated annual records were available. Predefined International Classification of Diseases (ICD)-10 codes for infection (A00-B99, G00-G09, I00-I02, J00-J22, L00-L08, O75.3, O85, O86, O98, P35-P39) were used to identify infection-related deaths.2 The dataset was also manually scrutinised to ensure complete ascertainment. Deaths were considered to be ‘infection-related’ if there was any mention of an infection on the death certificate, whether primary or contributory. Similar to the original analysis, children with no underlying medical condition recorded on the death certificate were considered previously healthy. Underlying comorbidities were manually categorised into major risk groups as in the 2003–05 analysis.2 Neonates, defined as aged <28 days on the date of death, were excluded due to differences in epidemiology, aetiology, risk factors, underlying conditions and coding of deaths. Where reported, infection-related deaths were categorised by their focus of infection; otherwise, they were considered as systemic infection without a reported focus. Infection-related deaths were also categorised by pathogen group (bacterial, viral, fungal, other) and, if recorded, the responsible pathogen. Prematurity was defined as live birth <37 weeks’ gestation. The age groups were categorised as postneonatal infants (aged 28 days up their first birthday), young children (aged 1 year up to their 5th birthday) and older children (aged 5 years up to their 15th birthday). Age-specific incidence was calculated by dividing total number of deaths and infection-related deaths over the 3-year period by the Office for National Statistics (www.statistics.gov.uk) mid-year population estimates for England and Wales during 2013–15 combined. Results from the current analysis were compared with the 2003–05 analysis.2 Poisson regression models were used to find age-specific and gender-specific incidence rate ratios (IRR), calculated with 95% CIs. Because errors were noted in the incidence estimates for overall and infection-related childhood mortality rates in the previous publication, these were recalculated for the current analysis. Variables that did not follow a normal distribution were expressed as medians and IQRs and compared using the Mann-Whitney U test. Categorical variables were reported as frequency of distribution or rates and expressed as 2×2 tables. Proportions were compared using the χ2 test or Fisher’s exact test, as appropriate.
Results
There were 5088 death registrations (annual incidence, 17.6 deaths/100 000) in children aged 28 days to <15 years in England and Wales during 2013–15 compared with 6897 during 2003–05 (23.9/100 000); equivalent to a 26% decline (IRR 0.74, 95% CI 0.71 to 0.77) over the decade; the proportional decline was similar across the age groups (table 1A). During 2013–15, there were 951 (18.7%, 951/5088) death registrations identified as infection-related compared with 1368 (19.8%, 1368/6897) during 2003–05, equivalent to an infection-related mortality rate of 3.3/100 000 compared with 4.8/100 000 during the two periods, respectively. This equated to a 31% decline (IRR 0.69, 95% CI 0.64 to 0.75) in infection-related mortality rate, with reductions of 45% in infants and 50% in young children, but a 22% increase among older children (table 1B). Infection-related deaths were higher in winter than in summer for children with and without comorbidities, which was also noted in 2003–05 (data not shown).
(A) Overall childhood deaths with childhood mortality rates and (B) gender-specific infection-related deaths with infection-related mortality rates by age-group in England and Wales during 2003-05 and 2013-15.
During 2013–15, the median age of children with an infection-related death was 1.3 years (492 days; IQR 100–2549 days); 40.7% (387/951) of deaths occurred in postneonatal infants, 26.4% (251/951) in young children and 32.9% (313/951) in older children. The contribution of infection-related deaths to total deaths was 16.8% (387/2310), 20.2% (251/1243) and 20.4% (313/1535), respectively. This compares with 21.4%, 27.3% and 12.3%, respectively, during 2003–05.
Comorbidities
Comorbidity was recorded in 55.0% (523/951) of death registrations during 2013–15 compared with 49.8% (676/1357) in 2003–05 (p=0.02). Children with comorbidity were older than healthy children (table 2). Comorbidity was recorded in 48.0% (186/387) of postneonatal infants, 54.2% (136/251) of young children and 64.2% (201/313) of older children. This compares with 49.2%, 43.7% and 61.3%, respectively, during 2003–05. The reported comorbidities varied with age. Prematurity (92/186, 49.5%) was more common in postneonatal infants, whereas chronic neurological conditions were more prevalent in young and older children (table 3).
Comparison of healthy children and those with underlying comorbidity who died of an infection in England and Wales during 2013–15
Distribution of comorbidity by age group in children with an infection-related death during 2013–15
Focus of infection
Where recorded (n=876), respiratory tract infection (RTI) was the most commonly reported focus (374/876, 42.7%) during 2013–15, while 28.8% (252/876) had systemic infection without a focus. Central nervous system (CNS) infections accounted for only 4.8% (42/876) (table 2). In 2003–05, almost half the infection-related deaths (629/1357, 46.4%) did not have a focus reported, while 30.1% (409/1357, 30.1%) had RTI and 16.1% (219/1357, 16.1%) had CNS infections. Systemic infections without a focus were more commonly identified in children without comorbidity (32.3%, 108/334) (p<0.01), whereas RTI was more frequently reported in children with comorbidities, mainly in postneonatal infants (p<0.01) (table 2).
Pathogens responsible
A pathogen was recorded in 63.0% (599/951) of infection-related deaths, similar to 2003–05 (837/1357, 61.7%). Pathogen reporting was more likely for children with comorbidity than healthy children (374/520 (71.9%) vs 225/431 (52.2%); p<0.05). This contrasts with 2003–05, when a pathogen was recorded for 50.3% (340/676) of children with comorbidity and 69.3% (472/681) of healthy children. Where recorded, 63.1% (378/599) of infection-related deaths were associated with a bacterial infection, 34.2% (205/599) viral and 2.5% (15/599) fungal infection in 2013–15; other aetiologies (1/599, 0.2%) were rare. This compares with 60.6% (507/837), 30.6% (256/837) and 7.8% (65/837), respectively, in 2003–05. Children with comorbidity had a higher proportion with recorded bacterial infections compared with healthy children (244/378 (64.6%) vs 134/378 (35.4%); p<0.05).
Bacterial infections
Of the bacterial infection-related deaths, the responsible pathogens varied by age and comorbidity status (tables 2 and 4A). Gram-negative infections were reported more frequently in children with comorbidity than healthy children (59.7% (37/62) vs 45.5% (50/110); p=0.07), but this was not statistically significant (table 4A). In healthy children, group B Streptococcus (GBS) was an important cause of death in postneonatal infants, while group A Streptococcus (GAS) was more common in children older than 1 year. Potential vaccine-preventable infection-related deaths included pneumococcal (n=12) and meningococcal (n=16) disease and occurred across all the age groups, mainly in healthy children. Their contribution to childhood deaths was much higher during 2003–05, with 138 meningococcal and 89 pneumococcal deaths recorded, accounting for 27.9% and 18.0%, respectively, of bacterial infection-related deaths where a pathogen was reported; GAS (16 cases, 3.2%), on the other hand, was less frequently recorded on death certificates during 2003–05.
(A) Bacteria and (B) viruses recorded by age-group in children who died of an infection in England and Wales during 2013-15
Viral infections
A viral cause of death was more frequently recorded in healthy children compared with those with comorbidities (89/225 (39.6%) vs 116/374 (31.0%), p<0.05) (table 2) and included respiratory syncytial virus (RSV, 28/205, 13.7%), adenovirus (26/205, 12.7%), influenza (20/205, 9.8%) and cytomegalovirus (CMV, 14/205, 6.8%) (table 4B), similar to 2003–05. The responsible viruses varied with age and comorbidity status. Most RSV-related deaths were in children with comorbidities, especially in the younger age groups; there were also RSV-related deaths in healthy infants. Adenovirus-related deaths occurred across all age groups, affecting both healthy children and those with comorbidities, and increased with age. Influenza-related deaths were also recorded in both groups and across all ages, particularly in the older children and those with comorbidities.
Other infections
Deaths due to pathogens other than bacteria and viruses were rare and nearly all due to fungal infections in children with comorbidity.
Discussion
Using electronic death registration records for England and Wales, we found that all-cause and infection-related childhood mortality rates beyond the neonatal period declined by 26% and 31%, respectively, during 2013-15 compared with 2003–05. Infections continued to contribute to around one in five childhood deaths. Large reductions were observed in postneonatal infants and young children. In older children, where comorbidity was recorded more frequently than younger age groups, the contribution of infections to deaths was greater than when compared with 2003–05. Half the infection-related deaths occurred in children with comorbidities and the most common focus was the respiratory tract. The responsible pathogen was recorded in 63.0% of death registrations and nearly two-thirds had a bacterial cause. Large reductions in meningococcal and pneumococcal deaths were observed, while deaths due to GAS were more frequent in 2013–15 than 2003–05, especially in toddlers and older children. Among viral infections, RSV, adenovirus, influenza and CMV remained the main responsible pathogens.
The lower all-cause mortality rate in 2013–15 compared with a decade previously is consistent with published reports, although the UK still has one of the highest childhood mortality rates among industrialised countries.3 5 A similar decline in the infection-related childhood mortality rates is, however, reassuring, although infection continues to contribute to one in five childhood deaths. There are no reliable data on comorbidity prevalence in the community but this is likely to be in the region of 5%–10%. If, as our analysis shows, half the infection-related deaths occur in children with comorbidity, then it follows that infections contribute to a larger proportion of deaths (10-fold to 20-fold higher) in children with comorbidity compared with healthy children. This is consistent with reports of higher all-cause mortality in children with comorbidity compared with healthy children.3 Notably, the proportion of children with a chronic condition who die continues to increase.3 4 6 7 In our cohort, the higher infection-related mortality rate in older children, who were also more likely to have a comorbidity, may be a consequence of better survival of younger children with comorbidities.
The contribution of the different pathogens to death remains similar across the decade. In healthy children, there were large reductions in pneumococcal and meningococcal deaths. The UK implemented a childhood pneumococcal immunisation programme in 2006, which led to large declines in childhood pneumococcal disease.8 9 The reduction in meningococcal deaths reflects the declining disease incidence across Europe since the early 2000s, most likely because of secular trends.10 In September 2015, the UK implemented meningococcal B vaccination into the national infant immunisation programme and, in August 2015, an adolescent meningococcal ACWY vaccination programme was also implemented to provide direct protection for teenagers and, over time, indirect (herd) protection across the population.11
GAS has emerged as a major pathogen responsible for bacterial-related deaths during 2013–15, reflecting a sharp increase in disease incidence since 2014 and reaching 33.2 cases/100 000 person-years by 2016, the highest rate in almost 50 years.12 While no vaccines are currently available for GAS, 15%–30% of cases are associated with varicella infection. Countries that have implemented varicella vaccination into their national childhood immunisation programmes have observed reductions of up to 70% in varicella-associated invasive GAS infections.13 Varicella vaccination is currently not included in the UK national immunisation programme,14 but could potentially reduce the burden of chickenpox and GAS infections and deaths in children. A continuing concern is the high rate of infection-related deaths due to enteric Gram-negative bacteria, particularly in immunosuppressed children. With increasing global concerns about multidrug resistance and lack of new antibiotic classes in the pipeline, the development of effective vaccines must be prioritised.
Among viral infections, influenza-related deaths occurred across all age groups. In the UK, the live attenuated influenza vaccine has been recommended for those aged 2–17 years and is currently offered to toddlers and primary school-age children, providing direct and indirect protection across all age groups.15 The small reduction in childhood influenza-related deaths from 30 cases during 2003–05 to 20 in 2013–15 could be due to the programme, but numbers are too small for formal statistical analysis. In addition to currently available vaccines,16 investigational vaccines that have the potential to reduce infectious disease burden and associated deaths, include the antenatal GBS vaccine,17 and several RSV vaccines across the life-course.18
Strengths and weaknesses
The strength of this methodology is that completion of a death certificate is a legal requirement. The consistency of using the same data source also allows for comparison over time. Analysing all death records for any mention of an infection also captures the entire burden of infection-related deaths, irrespective of whether an infection was recorded as the primary cause or contributed to the death on the death certificate, especially in those with comorbidities. Additionally, using annual death registration records is likely to have improved ascertainment because delayed registrations would also have been captured. The information contained within death certificates, however, is limited, may be incomplete and can vary over time, as discussed previously.2 Therefore, our estimates and the contribution of different pathogens must be considered a minimum. There were also methodological differences between the 2003–05 data, which were extracted using pre-specified ICD-10 codes for infection, and the 2013–15 data, which were also manually scrutinised.2 This, however, would overestimate the contribution of infection to childhood deaths during 2013–15. Finally, we are unable to compare our findings with other industrialised countries because of lack of similar data in the literature.
Implications and conclusions
Our results provide an evidence-base to prioritise research and recommendations to reduce infection-related childhood mortality in the UK. Our estimates are higher than the 2014 RCPCH report,3 perhaps because we included all deaths with any mention of an infection even if an underlying comorbidity was recorded as the primary cause of death. In addition to the priorities set in the report, such as configuration of health services with better integration of primary and secondary care, specific recommendations for preventing infections and improving the management of children with infections are required. These could include improving vaccine uptake, especially in hard-to-reach communities, implementing new vaccines, increasing education and awareness among parents and healthcare professionals, more robust evidence-based guidelines, optimising antibiotic prescribing, evaluating and addressing social determinants of infection-related deaths and ongoing surveillance to monitor changes in childhood infection-related deaths over time.
Footnotes
Twitter @shamezladhani
Contributors SL conceived the work; LF-A, IOO, GO and SL extracted and analysed the data. LF-A wrote the first draft of the manuscript. All authors contributed to the discussion and edited the manuscript. All authors approved the final submitted version.
Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval Public Health England has legal permission, provided by Regulation 3 of The Health Service (Control of Patient Information) Regulations 2002, to process patient confidential information for national surveillance of communicable diseases.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.