Epidemiological trends and terminology shifts

We observed an increase in deaths over time across the ABD case classification thresholds. Use of the term ‘excited delirium’ was more frequent at the beginning of the study and declined in use particularly after 2011. Although neither term is universally accepted, this shift corresponds to general preference for the term ABD over ‘excited delirium’. ^{Reference McGuinness and Lipsedge3} The increased citation of ABD as a cause of death in the latter period of the study is in contradiction to guidance issued in 2020 by the Forensic Science Regulator advising coroners to move away from use of either term as a stand-alone cause of death and to instead use these terms to reflect psychological and physiological disturbance alongside a more definitive cause of death (e.g. ‘amphetamine intoxication with acute behavioural disturbance’). ²⁴

Although classification of ABD cases using FFLM and RCEM frameworks represented a more standardised method for identification of the clinical signs and symptoms of ABD, the finding that all cases in this study met RCEM criteria indicates that there is limited specificity within this framework. Appropriate professional guidance could therefore be to use the term ABD only in cases in which two or more FFLM core symptoms are present, as proposed by a recent Delphi process. ^{Reference Humphries, Kelly, Sadik, Walker and Smith9} Application of this higher threshold would be likely to improve diagnostic consistency and overapplication of the term; indeed, in this study, only around one-third of cases met these diagnostic criteria, which demonstrates a need for clearer criteria in both clinical and forensic contexts.

Clinical features of ABD

Whereas the clinical presentation of cases (e.g. agitation, physical overactivity, non-compliance with law enforcement, autonomic dysregulation) was highly consistent with the diagnostic frameworks used, the method used was in part self-fulfilling, as cases in which ABD was not cited as a cause of death were identified in on the basis of the described symptomatology’s matching the diagnostic criteria. However, the prevalence of the different symptoms was similar across the different case types, indicating validity of the method used. Furthermore, the consistency in the described symptomatology across the ABD classification thresholds used supported the hypothesis that ABD may represent a distinct and recognisable clinical syndrome, albeit one with considerable diagnostic overlap and subjectivity. However, the frequent description of ‘bizarre’ behaviour – a symptom according to RCEM but not FFLM scoring criteria – suggests that a review of the current diagnostic tools may be needed for suitability and completeness.

Toxicological profile of ABD cases

The toxicological profile of ABD cases was strikingly different from that of cases from the wider NPSUM database. Stimulants were present in the majority of cases, with cannabis and dissociative drugs also more commonly detected in these cases compared with the overall cohort. By contrast, alcohol, opioids, sedatives and psychiatric medications were less frequently identified. These findings are consistent with those of previous case series and studies linking stimulant use to the development of ABD symptomatology ^{Reference McGuinness and Lipsedge3,Reference Ruttenber, Lawler-Heavner, Yin, Wetli, Hearn and Mash4,Reference Penders, Gestring and Vilensky17–Reference Fishbain and Wetli20} and further reinforce the role of multi-stimulant and illicit substance use in destabilising behavioural and physiological regulation. ^{Reference Kaye and Darke25–Reference Zawilska27}

Earlier work suggesting that cocaine concentrations in ABD deaths were lower than in other cocaine-related deaths ^{Reference Stevenson and Tracy28} has been used to argue that environmental factors such as restraint may be more important in the mechanism of death than drug toxicity. ^{Reference McGuinness and Lipsedge3} By contrast, we found that median cocaine and BZE concentrations were higher in ABD cases compared with all NPSUM deaths involving cocaine. However, interpretation of post-mortem cocaine concentrations is fraught with difficulty, as there is no consistent difference in concentrations between deaths in which cocaine was deemed causative and those in which its detection was deemed an incidental finding. ^{Reference Rooney, Sobiecka, Rock and Copeland29} Although detection of drugs may be useful in characterising cases of ABD, concentration data are unlikely to enable determination of individual toxicity thresholds.

Restraint as a mechanism of death

In agreement with previous studies on ABD, ^{Reference Martín-Ayuso, Pajuelo Castro, Santiago-Sáez, Herreros and Pinto Pastor10–Reference Humphries and Walker12,Reference Weedn, Steinberg and Speth15,Reference Vilke, DeBard, Chan, Ho, Dawes and Hall30} physical restraint prominently featured in the presented cases and was directly implicated as a factor contributing to death in more than half of the incidences in which it was used. The high proportion of cases in which ‘sudden collapse’ was described, particularly among cases in which restraint was used, provides support to theories that ABD is associated with a hyperadrenergic state that can lead to cardiovascular collapse ^{Reference Weedn, Steinberg and Speth15,Reference Steinberg16} – a state that would only be exacerbated by stimulant use. ^{Reference Coffin and Suen31,Reference Lange and Hillis32}

In many of the presented cases, ABD began at home or in public. Not all cases in which ABD was listed as a cause of death occurred in hospital, despite the clinical utility of the term being to signal urgently for emergency healthcare intervention. ^{Reference Humphries, Kelly, Sadik, Walker and Smith9} In many cases, there was failed community management or law enforcement involvement – often including restraint – before the delayed but eventual engagement with emergency services.

These findings further highlight previously voiced concerns regarding the use of restraint in suspected ABD cases ^{Reference McGuinness and Lipsedge3} and the need for alternative de-escalation strategies and immediate engagement with emergency healthcare, as has already been recommended in national clinical guidance. ^33,34 When ABD is suspected, earlier transfer to healthcare-led management pathways, including rapid tranquilisation, temperature control and cardiorespiratory monitoring, may reduce reliance on prolonged restraint and improve outcomes. It is prudent to reflect on whether inclusion of criteria such as ‘sustained non-compliance with police or healthcare staff’ by the RCEM – a scenario which often precedes restraint – risks defining ABD cases partly by the presence of restraint itself.

Demographic risk factors and health inequities

In comparison with the broader NPSUM cohort, decedents were younger and more likely to be male and living in areas of higher socioeconomic deprivation, and more likely to have a recorded minority ethnic background, although interpretation of ethnicity comparisons was limited by the high proportion of unknown ethnicity in the wider NPSUM cohort. Notably, ethnicity seemed to be recorded more frequently in ABD cases than in the wider comparator cohort, which may itself reflect differences in how these cases were perceived, described or documented. The overrepresentation of Black and Arabic individuals among cases with recorded ethnicity – a finding consistent with recent research using electronic health records, in which Black individuals were more likely to be described as having presented with ABD ^{Reference Polling, Das, Ariyo, Creary and Smith8} – is of particularly concern, as these findings do not reflect the ethnic make-up of the UK population, nor do they align with population-level patterns of drug use, in which stimulants are reportedly more commonly used by White people. ^{Reference Nutt, King and Phillips35,36} However, given the extent of missing ethnicity data, these findings should be considered exploratory. They suggest differences in the recognition, documentation and management of ABD and warrant further investigation using data-sets with more complete ethnicity recording.

Although data on mental health history were limited, the majority of those with available records had pre-existing psychiatric conditions, supporting calls for improved dual diagnosis care and early intervention strategies. ^34,37

Limitations

There are several limitations to this work that must be acknowledged. First, the NPSUM receives reports voluntarily from coroners. Therefore, not all deaths that met the described ABD classification thresholds from England, Wales and Northern Ireland may have been reported. Furthermore, we were only able to identify further ABD cases in which case narratives describing the behaviour of the deceased before death were provided. The total number of ABD cases is therefore likely to have been underestimated. Although the trends identified in this study are probably reflective of the true overall trends in ABD cases, it is possible that further themes and differing trends would emerge from cases that were not reported to the NPSUM, or from those reported with little detail.

Second, toxicological analyses were conducted in multiple UK toxicology laboratories with varying methods, extraction techniques and calibration ranges – assay details that are not routinely reported in NPSUM submissions. Therefore, although all laboratories were accredited by UKAS (or international equivalents), interlaboratory variability (e.g. in use of screening libraries and protocols) cannot be excluded.

Third, deprivation was measured using area-level Indices of Multiple Deprivation derived from the postcode of the decedent’s usual residence, rather than individual-level socioeconomic circumstances. Residence in a more deprived area does not necessarily mean that a person was themselves socioeconomically deprived, nor does residence in a less deprived area preclude individual disadvantage. Rather, this measure should be interpreted as a proxy for the broader social and environmental context in which individuals lived, including greater potential exposure to structural inequalities, poorer health outcomes and other deprivation-associated risk factors.

Finally, fatal ABD has been reported to occur in the absence of drug use. ^{Reference Kennedy and Savard38,Reference Pollanen, Chiasson, Cairns and Young39} In this study, however, it was only possible to identify fatal ABD cases from those reported to the NPSUM, which by definition require the decedent to have died in a context associated with psychoactive drug use.

Implications for future guidance

This study adds to the growing body of literature identifying ABD as a high-risk clinical presentation frequently associated with stimulant use, physical restraint and sociodemographic disadvantage. Although there remains a need for clearer diagnostic thresholds – such as requiring multiple FFLM criteria to ensure consistent application of the term ABD in forensic contexts – the consistent symptomatology and patterns of fatality evident in the cases presented here suggest a recognisable clinical syndrome with plausible shared pathophysiological mechanisms. The increase in ABD deaths over time and the overrepresentation of young and deprived ethnic minorities indicates a need for critical, equity-informed approaches to recognition and treatment of ABD. The frequent use of physical restraint and its contributions to sudden death warrant urgent reconsideration of current emergency response protocols. Rather than functioning as a stand-alone cause of death, ABD may be most useful as a time-critical clinical descriptor signalling severe agitation with physiological compromise that requires urgent medical assessment, rapid sedation where appropriate, active cooling, cardiovascular monitoring and minimisation of prolonged physical restraint through early healthcare-led de-escalation strategies. Future guidance should therefore focus on standardised recognition pathways and evidence-based emergency management rather than diagnostic labelling alone.