Temporal trends in time to asystole in patients considered for organ donation after circulatory death

Abstract

Introduction:

Prolonged time to asystole is a common barrier to successful organ donation after circulatory death, yet there have been few attempts to evaluate trends in its occurrence over time. This study investigated temporal trends in time to asystole among patients considered for donation after circulatory death in the UK (2014–2024).

Methods:

This observational study was conducted using NHS Organ Donation & Transplant (ODT) datasets. We analysed temporal trends in time to asystole and donation outcomes using descriptive statistics. Additionally, to explore evident impacts of the COVID-19 pandemic we performed pre-post comparisons and an interrupted time series (ITS) analysis.

Results:

Of 9466 participants, 58.5% proceeded to solid organ donation. Non-proceeding cases were primarily due to prolonged time to asystole (55.1%). Donation rates increased from 50% to 60% pre-2020 to consistently above 60% post-2022, whilst cases not proceeding due to prolonged time to asystole decreased from 25% to 20%. ITS analysis confirmed a significant post-pandemic reduction in time to asystole and a sustained drop in referrals, suggesting improved donation efficiency. Successful donors typically received higher levels of support and had lower neurological status compared to those with prolonged time to asystole.

Discussion:

The findings demonstrate reduced prolonged time to asystole occurrence and increased organ donation rates. Whilst there has been ongoing improvement over the study period there is evidence of a systemic change following the COVID-19 pandemic, which resulted in a smaller, more efficient referral pool. This may be attributed to better identification of suitable donors during the referral process, and while causation cannot be definitively established, the study supports ongoing evaluation and refinement of the donation pathways.

Keywords

euthanasia passive (E02.760.952.500)heart arrest (C14.280.383)time (G01.910)trends (Q000639)donor selection (E04.936.537.500)

Introduction

Background

Donation after Circulatory Death (DCD) is an increasingly important component of organ donation systems around the world, with many well established programmes for both controlled and uncontrolled DCD.^1,2 Controlled DCD almost exclusively refers to circulatory death that occurs following the withdrawal of life-sustaining treatment (WLST), most commonly in an Intensive Care Unit (ICU) setting. Following WLST, there is a variable period of deteriorating physiology, during which time organs can suffer from warm ischaemia. If this warm ischaemic time is prolonged, the organ damage incurred can be irreversible rendering the organ unsuitable for transplantation. This is termed prolonged time to asystole (PTA) and, when this occurs, organ donation does not proceed.³

The UK has a large DCD programme representing 56.6% of the 1713 patients for whom donation consent/authorisation was given in the 2024–2025 period.⁴ However, only 72.6% of these potential DCD donors proceeded to organ donation with the majority of cases halted due to PTA. The consequences of PTA are significant and include increased emotional distress for families⁵ in addition to resource implications for the service. Mobilisation of retrieval teams prior to WLST is required to ensure minimal warm ischaemic time, but in cases where PTA occurs, the utilisation of the resource without any organs suitable for transplantation potentially delays the process for other potential patients.

The implementation of PTA thresholds sees substantial variation with the time waited by retrieval teams ranging from 1 to 4 h across Europe² dependent on a variety of factors, such as clinical preference and utilisation of reperfusion techniques. The varying susceptibility to ischaemic damage observed across different organs can result in different thresholds for organ retrieval^6,7 depending upon the organs considered for donation, and the retrieval services available. These thresholds may also be impacted by the introduction of new techniques. For example, the increasing use of normothermic regional perfusion (NRP) allows for assessment of ischaemic damage prior to organ transplantation,⁸ and may lead to changes in the warm ischaemic time thresholds in cases where individualised organ damage can be adequately assessed. Despite the importance of time to asystole in the DCD donation process there has been limited description of the distribution of these times in consented/authorised donors or the trends in them over the years.

Objectives

We aimed to investigate the temporal trends in time to asystole within the UK DCD system that may help inform the structure of DCD systems in the UK. We also aimed to explore the trends in some of the limited physiological variables available over the same 10-year period.

Methods

Study design

This observational cross-sectional study of National Health Service (NHS) organ donation and transplant (ODT) DCD data was conducted and reported following the strengthening the reporting of observational studies in epidemiology (STROBE) guidelines.⁹

Setting, sample size and participants

The study population was all patients in UK hospitals for whom DCD consent/authorisation had been provided as available within the NHS ODT datasets.¹⁰ This dataset extract extended from 01 August 2014 until the most recently available records at the time of extraction, 31 March 2024.

Eligible participants were required to have an identifiable time from WLST to asystole. Participants with a time to asystole (TTA) that was negative or extremely long (>1 month) were excluded as these scenarios are likely the result of errors in data entry. Participants under 16 years of age were excluded.

Data sources

The data were provided by NHS Blood and Transplant with the variables requested extracted from the UK Transplant Database, which is their longstanding database of all donors, transplants and transplant patients in the UK. Where necessary data were also extracted from the Potential Donor Audit (PDA)¹¹ which comprises data gathered for each death in a critical care unit in the UK. Additionally, a field tabulating the daily number of DCD referrals was generated from the PDA to provide background detail on the flow of referrals prior to consent/authorisation.

The study was approved by the statistics sub-group of the National Organ Donation Committee of NHS Blood and Transplant, which undertakes review of information governance and public benefit for studies accessing anonymised healthcare data held by NHS Blood and Transplant.

Variables

The variables requested included the time of withdrawal, the time of asystole, primary diagnosis, cause of death, age, whether solid organ donation occurred and the reason for non-proceeding if the consented/authorised donation did not proceed.

Additionally, physiological variables recorded around the period of withdrawal were requested, including: mode of ventilation, spontaneous respiratory rate, spontaneous tidal volume, Glasgow Coma Scale (GCS), sedation status, alert/verbal/pain/unresponsive (AVPU) status (if sedated), vasoactive drug use, heart rate, systolic blood pressure, diastolic blood pressure, fraction of inspired oxygen (FiO2), and partial pressure of oxygen to fraction of inspired oxygen (PaO2/FiO2) ratio.

Selection, missingness and outliers

Many physiological variables recorded around the period of withdrawal had been recorded on multiple instances over the course of 12–48 h (i.e. are time-series, typically irregularly sampled). For these the latest non-missing value that was recorded prior to WLST was taken forward for analysis. Continuous physiological variables were visualised to evaluate outliers, and where outliers were identified thresholds for clinical validity were set (see Supplemental Figure 1 and Supplemental Table 1). Missingness for physiological variables is reported in Supplemental Table 6, with complete case analysis undertaken for the multivariable logistic regression, and with other statistical methods and visualisations resilient against missingness. The time to asystole was calculated as the difference between the timestamp of WLST and timestamp of asystole for all potential donors.

Statistical methods

For daily numbers of referrals and consents/authorisations, 10-day and 100-day centred moving averages were applied to reduce day-to-day variability while preserving underlying trends. These were plotted over time to visualise potential seasonal patterns and longer-term changes in the DCD referral pathway.

The proportion of patients proceeding to donation was calculated for monthly, quarterly and yearly periods. Mean yearly proportions were plotted with 95% confidence intervals, alongside scatter plots of monthly and quarterly proportions to visualise both short-term variation and long-term trends. Similarly, the proportion of patients not proceeding to donation specifically due to prolonged time to asystole was calculated using the number of non-proceeding cases attributed to PTA. These were plotted using the same approach as donation outcome proportions.

To examine changes in time to asystole across the study period, several complementary analytical approaches were used. Firstly, the cumulative proportion of patients reaching asystole within clinically relevant predefined thresholds (30, 60, 90, 120, 180, and 240 min) was calculated on a yearly basis. These thresholds were selected based on their clinical relevance to the retrieval of different organs, with shorter times (e.g. 30 min) being critical for cardiac donation and longer times (up to 240 min) potentially acceptable for kidney donation.

Secondly, box plots were generated to visualise the yearly distribution of time to asystole. For each calendar year, the interquartile range (IQR), median, and upper whisker (Quartile 3 + 1.5 IQRs) were calculated and plotted to show changes in the distribution over time. This approach was chosen as outliers are of particular importance in terms of resource utilisation. Additional box plots were created on quarterly and semi-annual bases to provide finer temporal resolution (Supplemental Figures 2 and 3).

For continuous variables (age, spontaneous tidal volume, heart rate, systolic and diastolic blood pressure, FiO2, and PaO2/FiO2 ratio), arithmetic means were calculated. For binary categorical variables (sedation status, and vasoactive drug use) a simple proportion of patients was calculated. Additionally, two continuous variables (spontaneous respirations, and GCS) were binarised due to observed profound skewing towards extremes, with spontaneous respirations of zero, representing apnoea, and a GCS of three, the maximum level of unresponsiveness. Finally, response level if sedated (recorded as AVPU) was treated as an ordinal variable ranging from zero to three (zero; alert, one; responsive to voice, two; responsive to pain, and three; unresponsive) allowing calculation of an arithmetic mean.

For each physiological variable, the latest recorded value prior to WLST was used. Both spearman’s correlation and multivariable logistic regression were used to evaluate the relationship between these, and PTA. Monthly and yearly means (median for AVPU) were calculated for the entire cohort, and separately for two key subgroups: patients who proceeded to donation, and patients who did not proceed due to PTA. These were visualised using scatter plots of both monthly and yearly mean values to allow visual comparison between successful donors and those experiencing PTA.

Finally, to specifically evaluate the impact of the COVID-19 pandemic, we undertook simple pre-post comparison analysis and a more complex interrupted time series (ITS) analysis. The pre-pandemic period was defined as prior to 2020/03/01 (the date of national lockdown in the UK) and the post-pandemic period as following 2022/03/01, and the analysis used weekly and monthly aggregated data. For binary outcomes (e.g. donation proportion, TTA thresholds), logistic generalised linear models (GLM) with a Binomial family were used, and for continuous count outcomes (e.g. referrals, consents), linear GLMs with a Gaussian family were used (Supplemental Equation 1). Each model estimated the change in level and slope during the pandemic transition and in the post-pandemic period, allowing comparison against a counterfactual trend (the projected trend had the pandemic not occurred).

All statistical analyses were performed using Python version 3.12.3 with specific packages and versions detailed in Supplemental Table 2. Figures were generated using the matplotlib (3.10.7) and seaborn (0.13.2).

Results

Participants

Of the total 10,389 consented/authorised participants identified, 9810 had an identifiable date and time of WLST, of which 9548 participants had an identifiable date and time of asystole. Of these, 61 participants were excluded as the time to asystole was negative, and a further 6 were excluded due to time to asystole values over 1 month. In total, this yielded 9466 participants for inclusion. This inclusion flow is visualised in Figure 1.

Figure 1.

Flow diagrams of inclusion and exclusion (left), and donation pathway (right).

The median age of the participants was 59 (range 16–85). Of the total 9466 participants included 5537 (58.5%) proceeded to solid organ donation, 3893 (41.1%) did not proceed to solid organ donation, and data were not available for the remaining 36 (0.4%). For all participants that did not proceed a justification was provided, typically prolonged time to asystole following WLST (n = 2145, 55.1%).

Main results

The number of participants referred and consent/authorisation provided are important context to consider when evaluating trends in the occurrence of prolonged time to asystole. Over the previous 9 years both the number of patients referred and DCD consent/authorisation provided remained relatively stable (Figure 2) with the notable exception of the start of 2020 when intensive care and organ donation services experienced substantial upheaval due to the COVID-19 pandemic.¹² There was an evident seasonal trend with rolling mean referral numbers commonly peaking over the winter period likely representing the background changes in mortality.¹³ This seasonality did not appear to impact the daily number of consent/authorisations provided for DCD.

Figure 2.

Daily referrals and consent/authorisation rate including centred 10 and 100 day rolling means for both to allow visualisation in trends whilst minimising noise.

Of the patients where consent/authorisation for DCD was provided, the proportion proceeding to donation was averaged over a range of time periods (Figure 3). Whilst the proportion proceeding to donation was relatively stable over the earlier years, with quarterly and yearly proportions between 50% and 60% until 2020, there was some improvement with all quarterly and yearly proportions since 2022 above 60%.

Figure 3.

Mean yearly proportion proceeding to donation with 95% CIs represented by the shaded area. Additionally scatter plots of the monthly and quarterly proportions are shown.

The proportion of patients consented/authorised for donation who did not proceed due to prolonged time to asystole is similarly visualised in Figure 4. The threshold for “‘prolonged time to asystole’ varies for different organs and will, therefore, be patient-specific based on the particular organs suitable for donation by the potential donor. Functional warm ischaemic time (FWIT) commences when systolic blood pressure falls below 50 mmHg or oxygen saturations less than 70% and continues until initiation of cold perfusion at the point of organ retrieval. A FWIT exceeding 30 min will render cardiac donation unsuitable where this time can reach up to 240 min for kidney donation. We find that the proportion of consented/authorised donors not proceeding due to PTA improved over time with a proportion around 25% being common prior to 2020 whilst in recent years this reduced to around 20%.

Figure 4.

Mean yearly proportion of consented/authorised patients not proceeding to donation due to prolonged time to asystole with 95% CIs represented by the shaded area. Additionally scatter plots of the monthly and quarterly proportions are shown.

Prior to visualising the trends in absolute time to asystole, the proportion of participants with asystole within clinically relevant predefined thresholds is visualised in Figure 5.

Figure 5.

Mean yearly proportions of time to asystole within clinically relevant thresholds. Note that the Y-axis (cumulative proportion) starts at 0.35 to improve legibility.

Given the large observed variation in time to asystole we visualised the distribution of time to asystole over a yearly basis using serial boxplots, as shown in Figure 6 with quarterly and semi-annual boxplots in the Supplemental Figures 2 and 3. This demonstrated a reduction in the time to asystole over the years with a particularly notable reduction in the third quartile and upper whisker values.

Figure 6.

Box plots of the yearly distribution of time to asystole with the box bounds representing the IQR (Q1–Q3), the internal line representing the median and the upper whisker representing Q3 + 1.5 IQRs.

Impact of the COVID-19 pandemic

To further investigate the notable changes observed around the COVID-19 pandemic, simple pre-post and ITS analysis was performed. The pre-post analyses can be found in Supplemental Tables 3 and 4, and the visualised ITS models, including the counterfactual trend, are shown in Figure 7.

Figure 7.

Monthly observed values (blue points) are shown with the fitted segmented regression model (red solid line) and the estimated counterfactual trend (green dashed line). Vertical dotted lines indicate the start (March 2020) and end (March 2022) of the COVID-19 transition period. Table of coefficients and associated p-values found in Supplemental Table 5.

Both analyses revealed a significant and sustained drop in the number of patient referrals, with the ITS demonstrating a significant drop in the level of referrals during the transition and post-pandemic, stabilising at a new level lower than both the pre-pandemic period, and that predicted by the counterfactual trend (p < 0.001).

Despite this reduction in referrals, the absolute number of monthly donations increased (median 47/month pre- vs 54/month post-pandemic, p < 0.001). This indicates a substantial increase in donation efficiency (donations per referral). Simple pre-post comparison showed the ‘donations per referral’ rate increased from 0.097 to 0.123 (p < 0.001), although the post-pandemic level was not statistically significant in comparison to the counterfactual trend in the ITS analysis.

To explore mechanisms for this increased efficiency we can also use simple pre-post testing to evaluate the improvement in TTA and PTA as previously identified. We observe statistically significant improvements in the proportion of patients reaching asystole within all tested time periods (p < 0.001). The ITS analysis was partially supportive of this with sustained and significant post-pandemic level change (p = 0.044) at 1 h, establishing a new, higher baseline. The simple pre-post comparison confirmed a significant drop in median TTA from 35 to 22 min (p < 0.001).

In contrast, while the simple ‘consents per referral’ rate improved (0.174–0.187, p = 0.005), the ITS model suggests this was largely a return to the pre-pandemic trend after a pandemic-related disruption. The post-pandemic level and slope for consent rate were not significantly different from the counterfactual (p = 0.661 for post-level), suggesting the primary driver of increased donation efficiency occurred downstream of the consent process, which would fit with the improvements in TTA.

Physiological analyses

Bivariate analysis and simple visualisation were used to guide variable handling for the multivariable logistic regression (Table 1), with both identifying several physiological variables prior to WLST as predictors of prolonged time to asystole. Notably, apnoea (indicating a lack of retained respiratory drive) was associated with significantly decreased odds of PTA, with similar findings for dependence on vasoactive drugs. The bivariate correlations with time to asystole are presented in Supplemental Figure 4. Visual inspection of the temporal trends in these variables (Supplemental Figure 5) indicates a shift in the physiological profile of consented patients over the study period, characterised by increasing proportions of patients with apnoea and vasoactive drug use, factors associated with shorter time to asystole. Trends separated by proceeding and not proceeding due to PTA are shown in Supplemental Figure 6.

Table 1.

Multivariable logistic regression analysis of variables associated with failed donation due to prolonged time to asystole.

Variable	OR (95% CI)	p-Value
Apnoea (no spontaneous respirations)	0.247 (0.208–0.294)	<0.001
Vasoactive drug use	0.483 (0.411–0.567)	<0.001
PaO2/FiO2 ratio	1.339 (1.251–1.433)	<0.001
FiO2	0.863 (0.808–0.921)	<0.001
Heart rate	0.887 (0.829–0.949)	<0.001
Age	1.091 (1.016–1.173)	0.017
Mean arterial pressure	1.042 (0.977–1.112)	0.212
Spontaneous TV	0.999 (0.942–1.060)	0.980
Sedation	0.999 (0.878–1.136)	0.986

Values are Odds Ratios (OR) with 95% Confidence Intervals (CI). p-values are derived from the logistic regression model. Predictors were standardised prior to analysis to give OR per standard deviation where appropriate. GCS was excluded due to high levels of missingness (56.2%), see Supplemental Table 6 for missing data analysis. Following this complete case analysis included a total of 6105 patients. Mean arterial pressure was used as the indicator for blood pressure, with systolic and diastolic blood pressures excluded due to the high levels of collinearity as seen in Supplemental Figure 4.

Discussion

Key results

We observed that among patients planned for DCD organ retrieval, the proportion successfully reaching solid organ donation has increased post-2022, and the proportion of patients not proceeding due to PTA has decreased. Our ITS analysis provides critical context, confirming that these improvements occurred alongside a significant and sustained drop in the total number of patient referrals post-pandemic. This divergence, with fewer referrals but stable or increasing absolute donations, points to a significant improvement in the efficiency of the DCD pathway, with the donation rate per referral increasing from ~10% to ~12%. This shift was accompanied by a clinically significant decrease in the average time to asystole and a potential change in the physiological profile of consented donors, who increasingly exhibited features associated with shorter time to asystole (i.e. apnoea and vasoactive drug use).

Interpretation

Interpreting these trends is complicated by the fact that the pandemic interruption occurred after a multi-year national strategy aimed at increasing referrals.¹⁴ This pre-pandemic ‘push’ created an upwards trend in referrals, and other donation metrics, with this trend forming the basis of the counterfactual in our ITS model. However, this trend was likely unsustainable, as the pool of potential ICU donors is finite limiting the improvement possible until reaching the point of saturation.

Acknowledging this potentially overambitious counterfactual is important for interpretation of our results, which must be considered in a broader context, and whilst bearing in mind the direct metrics as these are ultimately the values that represent the functioning of the organ donation system. For referrals, it may exaggerate the severity of the post-pandemic drop, with the observed new level potentially representing both a true pandemic-related change and a natural plateauing that was likely to occur anyway. Conversely, for efficiency metrics like ‘donations per referral’, this same ambitious counterfactual trend masks the observed improvement. While our ITS model did not find a statistically significant post-pandemic level change compared to the ever-increasing counterfactual (p = 0.171 for monthly), the simple pre-post data clearly shows a significant increase in efficiency (0.097–0.123, p < 0.001). This suggests the system is more efficient, even if it is not improving at the ambitious rate projected from pre-2020 data.

The most plausible explanation for these combined findings (fewer referrals, shorter TTA, and higher efficiency) is a systemic shift in the referral pool itself. The decrease in referrals is suggestive that the change may be mediated prior to specialist nurse involvement, likely at the level of the referring ICU teams. It is plausible that clinicians are now referring a smaller but ‘higher-quality’ cohort of patients who are clinically more likely to progress to asystole quickly. This may be due to several factors, including increased clinician experience and familiarity as DCD has become the more common mode of donation, leading to an improved ability to prognosticate. It may also reflect a shift in practice due to clinician workload and burnout, leading to a focus on donors with ‘clear potential’ rather than the ‘refer all’ approach promoted pre-pandemic, which may seem unrealistic due to their expectation of a limitation of retrieval resources. Alternatively, it is possible that withdrawal practices or the provision of palliative care have evolved in relation to the pandemic but this dataset does not capture these practices and so we cannot draw any conclusions here.

It is also possible that some of the observed trends in PTA could be explained by changes in retrieval practices. Normothermic regional perfusion (NRP) is a technique to restore the circulation to organs following circulatory arrest for the purpose of transplantation. NRP is used to allow recovery from FWIT, assessment of organ function/quality and facilitates a less hasty retrieval. Liver and kidneys transplanted following NRP have been demonstrated to have improved 12-month outcomes, and the technique is growing in popularity over the past decade. In 2019–2020, abdominal NRP was used in 5.7% of all UK DCD retrieval procedures,¹⁵ increasing to 20% in 2023–2024.¹⁶ As it allows for assessment of organs from the borderline donor who may have experienced a longer FWIT than would have previously been accepted, there may be a tendency to proceed with donation via NRP where previously PTA would have led to donation step down, although no formal protocols exist to this effect. While this observation may help to explain a drop in the number of non-proceeding donors due to PTA, it does not account for the associated reduction in time to asystole.

Generalisability and implications

The observed reduction in time to asystole and lower rates of non-proceeding donation due to PTA appear to be driven by a fundamental change in the donor referral population. It is possible that our observation of reduced time to asystole and lower rates of non-proceeding donation due to PTA could be at least partially accounted for by an increasingly risk-averse clinical team, only willing to proceed DCD donation in those most likely to progress. With increasing demands on retrieval teams, clinicians may inadvertently be introducing selection bias.

Although we lack physiological data for non-consented referrals, the trends in Supplemental Figure 5 show that the consented population has potentially been moving towards physiological characteristics associated with shorter time to asystole (i.e. vasoactive drug use, spontaneous respirations/apnoea). This could support the hypothesis that the reduction in total referrals reflects improved upstream triage, filtering out patients unlikely to proceed before the consent stage. However, without physiological data for non-referred or non-consented patients it is not possible to fully untangle this relationship which could represent more general changes in the admission or management of the overall ICU population.

As DCD organ utilisation remains a relatively new pathway in the UK, clinical experience and confidence is growing within the transplant community and potential recipient patient group. Early utilisation criteria were cautious for understandable reasons, however increasing evidence suggests that transplant outcomes remain good despite higher FWIT with good donor selection, even in the absence of use of NRP.^17
–19 As this evidence and experience grows, so PTA might be predicted to fall as a larger fraction of DCD donor organs are retrieved despite higher FWIT. Of note, FWIT and time to asystole are not the same as explained above, and although it is possible for short time to asystole to also include higher FWIT, in general we might expect the two measures to be reasonably well correlated. Our data show a clear pattern of reduction in time to asystole, suggesting it is a change to the potential donor population (in terms of physiology or end of life care received) which is primarily responsible, rather than altered management of patients with prolonged FWIT.

While our data shows referral numbers have dropped, the key is to ensure this reflects better selection rather than missed opportunities. It is important that there are robust mechanisms for feedback to referring teams regarding the success of donation, with specialist nurses for organ donation (SN-ODs) well placed to support this, given their established remit for education and liaison with critical care staff alongside their family-facing responsibilities.²⁰ This includes encouraging reflection on patients who are not referred but then go on to have a short time to asystole that may have been suitable for donation. Notably, recent audits suggest that the number of missed referrals may have been falling prior to the pandemic, lending further support to the idea of a more experienced and accurate selection process at the ICU level. This highlights the importance of regular and robust publication of audits to monitor these trends and pathways at a national level.

It should also be noted that although strengthening experience and pathways offer clear benefits, both should remain flexible enough to accommodate changing requirements regarding transplantation, with new techniques and technologies potentially widening the time to asystole windows that are appropriate. This is additionally an important consideration for any efforts to develop tools for predicting time to asystole of which there are several²¹ although none are currently used within the UK DCD system. In recognition of the challenge and importance of prognostication of time to asystole, NHS Organ Donation and Transplantation have recently launched a national project (Sustainability and Certainty in Organ Retrieval) aiming to standardise and improve assessment of time to asystole in the potential DCD donor.

Limitations

As a cross-sectional observational study there are limitations placed on our ability to attribute causation. Whilst we have demonstrated that over time there have been changes in the physiological variables in those with consent/authorisation, whether these changes reflect changes in the overall potential DCD population, or changes to selection patterns in who we are choosing to approach or progress to donation, cannot be determined due to the lack of physiological data for patients in whom consent/authorisation was not provided/requested. Given this, and the retrospective focus of the analyses, no predictive modelling was undertaken. Additionally, we were unable to evaluate potential changes in donor demographics during the time frame, with the exception of age, due to the limited variables available in dataset.

Finally, a specific limitation of the ITS analysis is the reliance on a linear counterfactual trend as discussed in the interpretation. As the pre-2020 surge in referrals was likely approaching saturation within a finite donor pool, the projected linear growth is likely an overestimation. Consequently, the counterfactual may exaggerate the post-pandemic drop in referrals while simultaneously masking the true magnitude of efficiency improvements. This necessitates careful nuanced interpretation rather than a focus on statistical significance and direct conclusions.

Conclusion

We observed an increase in the proportion of patients consented/authorised for DCD successfully proceeding to donation between 2014 and 2024 and a corresponding decrease in the proportion of non-proceeding donors due to prolonged time to asystole. The COVID-19 pandemic appears to have been an inflection point, with an increasingly efficient DCD process characterised by a reduced referral pool, stable donation numbers and shorter times to asystole. This somewhat paradoxical finding may be driven by improved donor selection at the referring-clinician level. Finally, we highlight some physiological variables and factors that appear to differentiate between successful donors and those experiencing prolonged time to asystole, such as spontaneous respirations and vasoactive drug use, that may warrant further exploration.

Supplemental Material

sj-docx-1-inc-10.1177_17511437261450006 – Supplemental material for Temporal trends in time to asystole in patients considered for organ donation after circulatory death

Supplemental material, sj-docx-1-inc-10.1177_17511437261450006 for Temporal trends in time to asystole in patients considered for organ donation after circulatory death by Craig Nicolson, Athanasios Tsanas, Dale Gardiner, Dan Harvey, Martin Shaw, Nazir Lone and Kathryn Puxty in Journal of the Intensive Care Society

Supplemental Material

sj-docx-2-inc-10.1177_17511437261450006 – Supplemental material for Temporal trends in time to asystole in patients considered for organ donation after circulatory death

Supplemental material, sj-docx-2-inc-10.1177_17511437261450006 for Temporal trends in time to asystole in patients considered for organ donation after circulatory death by Craig Nicolson, Athanasios Tsanas, Dale Gardiner, Dan Harvey, Martin Shaw, Nazir Lone and Kathryn Puxty in Journal of the Intensive Care Society

Footnotes

Acknowledgements

This work was conducted using data provided by NHS Blood and Transplant. The authors of this study would like to acknowledge and thank all of those at NHS Blood and Transplant who have contributed to the collection of the dataset and those who support its maintenance and provision to researchers.

ORCID iDs

Craig Nicolson

Nazir Lone

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supplemental material

Supplemental material for this article is available online.

References

Barboza

Dhanani

Browning

, et al. Trends in donation after circulatory determination of death donor utilization: Lessons from Houston. Transplant Rep 2023; 8: 100135.

Lomero

Gardiner

Coll

, et al. Donation after circulatory death today: an updated overview of the European landscape. Transpl Int 2020; 33: 76–88.

Suntharalingam

Sharples

Dudley

, et al. Time to cardiac death after withdrawal of life-sustaining treatment in potential organ donors. Am J Transplant 2009; 9: 2157–2165.

Statistics and Clinical Research, NHS Blood and Transplant. NHS ODT Potential Donor Audit Report 2024-25, https://www.odt.nhs.uk/statistics-and-reports/potential-donor-audit-report/ (2025, accessed 13 November 2025).

Manara

Murphy

O’Callaghan

Donation after circulatory death. Br J Anaesth 2012; 108: i108–i121.

Mihaylov

Mangus

Ekser

, et al. Expanding the donor pool with the use of extended criteria Donation After circulatory death livers. Liver Transplant 2019; 25: 1198–1208.

Reid

AWN

Harper

Jackson

, et al. Expansion of the kidney donor pool by using cardiac death donors with prolonged time to cardiorespiratory arrest. Am J Transplant 2011; 11: 995–1005.

Truog

Flescher

Ladin

Normothermic regional perfusion - the next frontier in organ transplants?

JAMA 2023; 329: 2123–2124.

von Elm

Altman

Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008; 61: 344–349.

10.

ODT Clinical–NHS Blood and Transplant. ODT Clinical–Statistics and Reports–Access Data. ODT Clinical–NHS Blood and Transplant, https://www.odt.nhs.uk/statistics-and-reports/access-data/ (2025, accessed 13 November 2025).

11.

ODT Clinical–NHS Blood and Transplant. ODT Clinical–Statistics and Reports–Potential Donor Audit Report. ODT Clinical–NHS Blood and Transplant, https://www.odt.nhs.uk/statistics-and-reports/potential-donor-audit-report/ (2025, accessed 13 November 2025).

12.

Plummer

Alcock

Madden

, et al. The impact of COVID-19 on organ donation and transplantation in the UK: lessons learned from the first year of the pandemic. Anaesthesia 2022; 77: 1237–1250.

13.

Brown

Fearn

Wells

Exploratory analysis of seasonal mortality in England and Wales, 1998 to 2007. Health Stat Q 2010; 48: 58–80.

14.

ODT Clinical–NHS Blood and Transplant. Taking organ utilisation to 2020. ODT Clinical–NHS Blood and Transplant, https://www.odt.nhs.uk/odt-structures-and-standards/key-strategies/archived-strategies/taking-organ-utilisation-to-2020/ (2013, accessed 13 November 2025).

15.

Statistics and Clinical Research, NHS Blood and Transplant. NHS ODT Annual Activity Report 2019-20, https://nhsbtdbe.blob.core.windows.net/umbraco-assets-corp/19481/activity-report-2019-2020.pdf (2020, accessed 13 November 2025).

16.

Statistics and Clinical Research, NHS Blood and Transplant. Annual Report on the National Organ Retrieval Service 2023-24, https://nhsbtdbe.blob.core.windows.net/umbraco-assets-corp/34308/annual-report-on-the-national-organ-retrieval-service-202324.pdf (2024, accessed 13 November 2025).

17.

Pasrija

DeBose-Scarlett

Keck

, et al. (192) prolonged warm ischemic time is safe for cardiac donation after circulatory death. J Heart Lung Transplant 2023; 42: S94–S95.

18.

Brann

Jackson

White

, et al. (191) I mpact of functional warm ischemic time on short term outcomes in donation after circulatory death heart transplantation. J Heart Lung Transplant 2023; 42: S94.

19.

Alghannam

Fine

Howard

, et al. Impact of warm ischemia time on donation after circulatory death kidney transplant outcomes. Clin Transpl 2024; 38: e15436.

20.

Tocher

Neades

Smith

, et al. The role of specialist nurses for organ donation: a solution for maximising organ donation rates? J Clin Nurs 2019; 28: 2020–2027.

21.

Nicolson

Burke

Gardiner

, et al. Predicting time to asystole following withdrawal of life-sustaining treatment: a systematic review. Anaesthesia 2024; 79: 638–649.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.03 MB

0.00 MB

1.51 MB