Article Text
PDF
Abstract
Objective We investigated sudden unexpected death in infancy (SUDI) autopsy data from 1996 to 2015 inclusive, comparing findings from infants with and without pre-existing medical conditions.
Design Large, retrospective single-centre autopsy series.
Setting Tertiary paediatric hospital, London, UK.
Methods Non-identifiable autopsy findings were extracted from an existing research database for infants older than 7 days up to and including 365 days old who died suddenly and unexpectedly (SUDI; n=1739). Cases were classified into SUDI with pre-existing condition (SUDI-PEC) (n=233) versus SUDI without PEC (SUDI non-PEC) (n=929), where PEC indicates a potentially life-limiting pre-existing medical condition. Findings were compared between groups including evaluation of type of PEC and whether the deaths were medically explained (infectious or non-infectious) or apparently unexplained.
Results Median age of death was greater in SUDI-PEC compared with SUDI non-PEC (129 days vs 67 days) with similar male to female ratio (1.4:1). A greater proportion of deaths were classified as medically explained in SUDI-PEC versus SUDI non-PEC (73% vs 30%). Of the explained SUDI, a greater proportion of deaths were non-infectious for SUDI-PEC than SUDI non-PEC (66% vs 32%). SUDI-PEC (infectious) infants were most likely to have respiratory infection (64%), with susceptible PEC, including neurological, prematurity with a PEC, and syndromes or other anomalies.
Conclusion SUDI-PEC deaths occur later in infancy and are likely to have their death attributed to their PEC, even in the absence of specific positive autopsy findings. Future research should aim to further define this cohort to help inform SUDI postmortem guidelines, paediatric clinical practice to reduce infant death, and to reduce the risk of overattribution of deaths in the context of a PEC.
- neonatology
- pathology
- infant
- child health
- mortality
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Sudden unexpected death in infancy (SUDI) remains one of the most common presentations of death in children under 1 year. Infants with pre-existing conditions (PEC) have a higher mortality rate. However, there are little data comparing causes of sudden unexpected deaths in apparently healthy infants (SUDI non-PEC) to those with such PEC (SUDI-PEC).
WHAT THIS STUDY ADDS
We have shown SUDI cases with and without PEC have different autopsy findings, including apparent attribution of death to underlying conditions, which will help to further understand SUDI in different patient cohorts.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The findings highlight the need to further define specific SUDI cohorts, to better understand SUDI mechanisms, which will inform postmortem examination guidelines, and guide paediatric clinical practice to reduce risk in those with PEC.
Introduction
Postneonatal infant mortality rates have largely plateaued in the UK since the major reduction with the ‘Back to Sleep’ campaign in the 1990s.1 From 2013 to 2019, there was virtually no change in the unexplained infant death rate at ~0.3/1000 live births.2 The UK has fallen behind European Union countries who have continued to show better progress with reducing both neonatal and under-five mortality rates.3
The most common overall causes of infant mortality in England and Wales are congenital abnormalities and immaturity-related conditions, with unexplained deaths accounting for 6.7% of all infant deaths.4 However, sudden unexpected death in infancy (SUDI) represents the most common presentation of postneonatal infant death.5 Such cases are all investigated, on behalf of HM Coroner, almost all of which undergo detailed postmortem examination, including ancillary investigations, according to a standard protocol, carried out by specialist paediatric pathologists. Nevertheless, around a half of such cases remain unexplained following investigation and are classified as unexplained SUDI (sudden infant death syndrome or cot death) depending on the specific circumstances.5
Sudden unexpected death may also affect children with underlying pre-existing medical conditions, many of which are known to have an impact on overall mortality rates, although this group is poorly studied.4 There have been some recent studies investigating the causes of death in older children over the age of 1 (sudden unexpected death in childhood-PEC); however, there are little data on autopsy findings from sudden unexpected deaths in apparently healthy infants (non-PEC) compared with those with pre-existing conditions (PEC).6
This study therefore examines data from a large, retrospective autopsy series investigating SUDI in a single centre from 1996 to 2015. All autopsies were performed at the Great Ormond Street Hospital for Children by specialist paediatric and perinatal pathologists under the guidance of HM Coroner, and the findings were systematically recorded, including results of each stage of the examination and ancillary investigations.
Methods
Non-identifiable data were extracted from the autopsy research database, containing all information from the time of the autopsy, including the clinical history and circumstances, as provided at the time of the autopsy. All autopsies were carried out according to a standard protocol in one centre. ‘Positive’ autopsy findings were those deemed sufficient to provide a likely cause of death, such as histological evidence of pneumonia, for example. ‘Negative’ autopsy findings were those without such evidence.
There were 1739 SUDI cases overall, defined as infants older than 7 days up to and including 365 days who died suddenly and unexpectedly, where the death was not reasonably expected to occur in the previous 24 hours. Forensic cases performed by forensic pathologists on behalf of the police, including probable non-accidental injury, homicide, and non-natural deaths such as trauma and accidental deaths, were excluded, resulting in 1162 SUDI cases from inclusion.
Cases were classified into SUDI-PEC versus SUDI non-PEC, where PEC indicates a documented, potentially life-limiting pre-existing medical condition. Non-PEC also included premature infants who were otherwise apparently healthy and those who spent time in hospital after birth but were discharged without sequelae. Premature infants were aged as per the documentation at the time of autopsy. Due to the limited information available, infants born to mothers with pregnancy complications that were otherwise healthy were deemed non-PEC. Cases were classified based on the information available at the time of the postmortem examination. Borderline cases were discussed among the research team (RC, NJS, JCH), and consensus was reached. Cases were further categorised into PEC subgroups. Comparisons were made between PEC and non-PEC cases in terms of age, sex and whether the deaths were explained by autopsy findings: infectious or non-infectious, or unexplained. For the purposes of this study, histological features of autopsy were not re-examined; the final conclusion of the reporting pathologist was used.
Two-tailed comparisons of two proportion Z-tests were carried out using p<0.01 as the significance level, since multiple comparisons were performed. The following categories were compared: age of death (PEC vs non-PEC), sex of infant (male vs female), explained SUDI-PEC versus non-PEC, explained SUDI-PEC (infectious) versus non-PEC, unexplained SUDI-PEC versus non-PEC, explained SUDI-PEC (infectious) respiratory infection versus other types of infection.
Patient and public involvement
Patients and the public were not involved in the design of this study.
Results
There were a total of 1162 SUDI autopsies during the period of 1996–2015 at the centre. Of these, 233 were SUDI-PEC deaths and 929 were SUDI non-PEC deaths. SUDI-PEC was further divided into condition subcategories (figure 1, table 1).
Pre-existing condition (PEC) subcategories including percentage occurrence.
Pre-existing conditions
Overall, 448 SUDI cases were explained (39%) and 714 SUDI cases were unexplained (61%). Of the explained SUDI cases, 247 (55%) were infectious and 201 (45%) were non-infectious (table 2).
Proportions of SUDI explained versus unexplained
The median age of all SUDI was 77 days. Median age of death was greater in SUDI-PEC than SUDI non-PEC (129 days vs 67 days, z=−8.359, p<0.001) (figure 2). There were more males than females overall (1.4:1) (p<0.001), with similar ratios for both the PEC and non-PEC cohorts (z=0.0372, p=0.97).
Age of death (days) in SUDI-PEC versus SUDI non-PEC with trend line. Arrows represent median age of death (days) for PEC and non-PEC. SUDI-PEC, sudden unexpected death in infancy with pre-existing condition.
A greater proportion of deaths were explained in SUDI-PEC than in SUDI non-PEC (73% vs 30%, z=11.9176, p<0.001) (figure 3, table 3). Of the explained SUDI, there were a greater proportion of non-infectious deaths in PEC than in non-PEC (66% vs 32%), and a greater proportion of infectious deaths in non-PEC than in PEC (68% vs 34%) (z=6.8941, p<0.001) (figure 3, table 3). Of the explained SUDI (infectious), respiratory infections were the most common across both PEC (64%) and non-PEC (38%) cohorts (table 4, figure 4). The following PEC contributed the greatest number of respiratory infectious deaths in the PEC cohort: neurological, prematurity with a PEC, and syndromes or other anomalies (online supplemental figure 1, online supplemental table 1). Of the 111 explained SUDI-PEC (non-infectious) deaths, there were 20 cases (18%), where the cause of death was not fully explained by autopsy, and the cause of death was assigned on the balance of probability to the underlying PEC. Of these 20 cases, the majority had cardiovascular PEC (75%).
Supplemental material
Supplemental material
Proportion of SUDI explained versus unexplained in non-PEC and PEC. SUDI-PEC, sudden unexpected death in infancy with pre-existing condition.
Proportions of SUDI-PEC versus SUDI non-PEC showing unexplained versus explained SUDI, including explained SUDI subcohorts
Types of infection as cause of death in SUDI explained PEC (infectious) versus SUDI explained non-PEC (infectious)
Types of infection as cause of death in SUDI explained PEC (infectious) versus SUDI explained non-PEC (infectious). PEC, pre-existing condition; SUDI, sudden unexpected death in infancy.
For the cohort of 62 infants with a cardiovascular PEC, 87% were explained SUDI (non-infectious). Just 5% of cardiovascular PEC cases were unexplained SUDI, compared with the all SUDI-PEC (unexplained) proportion of 27%.
Of the 58 cases of prematurity with a PEC, seven were considered neurological PEC, one of which classified as hypoxic ischaemic encephalopathy (death was explained non-infectious).
Discussion
The findings of this study have described, for the first time, the differences in presentation and autopsy findings between otherwise healthy infants presenting with SUDI and those who died suddenly and unexpectedly, but with an underlying known medical condition. The findings have demonstrated that an apparent medical cause of death is more likely to be given in those dying suddenly and unexpectedly with a PEC, even in the absence of specific positive autopsy findings, and where a superimposed medical cause of death is apparent, this is most likely to be infection, particularly of the respiratory tract.
The median overall age of death of 77 days is consistent with what is already known about the most vulnerable period being 2–4 months of age. The SUDI-PEC cohort presented with death at an older age than SUDI non-PEC, suggesting that the mechanisms for such deaths are likely to be different between these groups. We predicted that the age of death should be bimodal in the PEC cohort; the first peak representing those presenting with the same mechanism as other SUDI deaths with incidental PEC, and the second peak representing deaths related directly to the underlying PEC or other specific mechanism. However, no such age frequency distribution was observed, with the age of death spread throughout the range in those with PEC, further suggesting that the cause of death may have different mechanisms in the PEC and non-PEC cohorts. The greatest difference in frequency of death between PEC and non-PEC appears to be in the first 100 days of life. It is possible that some of those early deaths in the non-PEC cohort are potential PEC infants who are yet to be diagnosed; however, the majority of the PEC conditions are likely to be diagnosed antenatally or in early life, for example, cardiovascular, prematurity+PEC, syndromes, other anomalies. The ratio of 1:1.4 (females:males) is consistent with previous data and confirms that males are more predisposed to SUDI, regardless of presence or absence of PEC, although the mechanism remains unclear.7 8
A greater proportion of SUDI cases were apparently explained with a provided specific medical cause of death in SUDI-PEC versus non-PEC. We propose two possible explanations. First, there may genuinely be more explained PEC deaths due to their underlying condition and/or predisposition to particular complications or superimposed diseases. Second, the SUDI-PEC ‘explained’ cause of death may be artificially high since many otherwise ‘unexplained’ deaths following autopsy examination may be allocated an explained cause, with the assumption that this must be related to their PEC/predisposition. This is an important factor for understanding the epidemiology and mechanisms of death in such cases, and there is an area for future research in helping to further define the causes of death within the PEC cohort in relation to their underlying medical conditions.
Similarly, within the explained SUDI-PEC deaths there were a greater proportion of deaths which were categorised as non-infectious compared with explained SUDI non-PEC deaths. This could again be due to a greater proportion of the SUDI-PEC cases being explained by their underlying condition, and therefore more readily categorised as non-infectious. Actual rates of infectious deaths are likely similar or even greater among SUDI-PEC cases, but denominator data are not available since the inclusion criteria for this study were that the death was initially regarded as unexplained by the clinical team and was hence referred to the Coronial system. Deaths which are not deemed to be sudden and unexpected, many of which may include infectious deaths in those with underlying PEC, would have a medical death certificate issued by the clinician, and therefore may not be referred to the Coronial system for further investigation.
Examination of the explained SUDI-PEC (non-infectious) reveals that those with negative autopsy findings but with the underlying PEC regarded as the cause of death were most likely to have cardiovascular PEC. Only 5% of cardiovascular PEC deaths were classified as unexplained, compared with 27% of all SUDI-PEC deaths. It is therefore likely that at least some of the apparently ‘explained’ cardiovascular PEC deaths represent unexplained SUDI, but it is impossible with current autopsy methods to ascertain a definitive cause of mechanism of death, such as whether an infant had a fatal arrhythmia, for example. HM Coronial autopsies are required to report all relevant positive and negative findings in addition to an overall likely cause or mechanism of death on the balance of probability. The authors recognise the inherent uncertainty in interpretation and designation of cause of death in autopsy studies. Further research, including development of novel autopsy biomarkers for specific mechanisms of death, may in future allow improved classification.
Our data show that those infants with a PEC, who die of an infectious cause, are most likely to die of a respiratory infection (figure 4, table 4). Respiratory infectious deaths are relatively over-represented in the PEC cohort compared with non-PEC (64% vs 38% of infectious deaths) cohort, especially in those with neurological conditions, prematurity with another PEC, and syndromes or other anomalies (online supplemental figure 1, online supplemental table 1). This finding is important for paediatric clinical teams to consider in terms of reducing such deaths, for example, by lowering respiratory infection treatment thresholds for infants with a PEC, and by developing improved detection for those at increased risk of respiratory infectious deaths.
The best known data to compare to PEC in infants in the general population were from the National Congenital Anomaly and Rare Disease Registration Service (online supplemental table 2).9 Their 2019 summary reports live births and deaths in infants <1 year with ‘congenital abnormalities’, defined to include structural, chromosomal and genetic conditions. While this is not entirely representative of the current PEC cohort, it represents the best available evidence for death rates in this group, since there are no directly comparable datasets available. The report indicates that 6.5% of live births with congenital abnormalities died in 2019 but no specific subclassifications for cause of death are available. Given that the current SUDI-PEC cohort was over-represented at 20% of the total SUDI cohort, a much greater proportion than the general population, this further suggests that PEC infants are significantly more susceptible to SUDI presentation.
Supplemental material
Review of the literature regarding previous SUDI autopsy studies that compare PEC and non-PEC is limited. Three studies that included comparison with a PEC were available (online supplemental tables 3 and 4). The present study is more than twice as large as all previous studies combined, but the data are consistent with others reporting that SUDI-PEC accounts for around 10–20% of all SUDI cases. One previous study included seven PECs as part of their multifactorial risk factor study (major structural birth defect, respiratory distress syndrome, intraventricular haemorrhage, necrotising enterocolitis, periventricular leukomalacia, patent ductus arteriosus and bronchopulmonary dysplasia)10; another study defined 17 infants with ‘previous medical history’, 15 of which had cardiac conditions and two had ‘other known conditions’11; and another study defined their SUDI cohort as <2 years of age, and reported that the most common cause of death in those with stable PEC was non-infectious; among the SUDI-PEC (infectious) cases, respiratory infections were again the most common cause of death.12 There is inherent subjectivity in the literature around classifying PEC; however, this study made careful distinction as to whether a pre-existing medical condition was potentially life limiting or not.
The present study is by far the largest globally to examine autopsy findings from SUDI cases with and without PEC. The findings show that SUDI-PEC deaths occur later in infancy, with a median age of 129 days compared with 67 days in non-PEC, suggesting a different cause of death mechanisms between the two. SUDI-PEC cases have a greater proportion of apparently medically explained deaths and are more likely to have their deaths attributed to their PEC, even in the absence of positive autopsy findings. Those with a cardiovascular PEC have a disproportionately small number of unexplained deaths, with most deaths presumed to be due to the underlying condition. Respiratory infectious deaths were relatively over-represented in the PEC cohort, with some PEC appearing more susceptible. Further research is needed to define the PEC cohort, along with collating more information premortem and postmortem. Future studies should describe SUDI and SUDI-PEC in a similar format to enable comparisons. The findings will help inform SUDI postmortem guidelines and paediatric clinical practice to reduce infant death, and to reduce the risk of overattribution of deaths in the context of a PEC.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
Existing research ethics committee approval was used for this study via the Great Ormond Street Hospital for Children NHS Foundation Trust (R&D reference: 22SH17).
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors Guarantor: NJS.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.