Treatment withdrawal and acute brain injury: an integral part of care

“The success of intensive care is not to be measured only by the statistics of survival, as though each death were a medical failure. It is to be measured by the quality of lives preserved or restored; and by the quality of the dying of those in whose interest it is to die.” G. R. Dunstan, Professor of Moral and Social Theology, 1995 [1] Critical care medicine has evolved rapidly over the last two decades, with therapeutic and technological advances leading to improved outcome in a wide variety of life-threatening conditions. This is particularly true for acute brain injury where improved understanding of pathophysiology, in association with advances in monitoring and imaging techniques, has led to the introduction of more effective and individualised treatment strategies that have translated into improved outcomes [2-4]. Acute brain injury can occur from a variety of causes including trauma, subarachnoid and intracerebral haemorrhage, and acute ischaemic stroke. Patients who were previously deemed to be unsalvageable are now being admitted to intensive care units, allowing many to recover beyond initial expectations [4, 5]. Others will have a poor outcome despite maximal intervention and it is essential that early aggressive treatment is linked to a timely switch to compassionate end-of-life care if a satisfactory degree of neurological recovery is not possible. There are important differences between brain-injured patients compared with a general population of critically ill patients. First, brain injury tends to occur suddenly, often in patients who were previously healthy and, in some cases, relatively young. Second, although the brain injury may result in poor functional outcome, progression to death is unlikely in many cases unless treatment is restricted or withdrawn. Finally, the majority of patients immediately become incapable of making their own decisions, leaving families to take on the role of surrogate decision makers at a time of great uncertainty. Acute brain injury is a heterogeneous condition with respect to pathophysiology and severity, as well as cause, and the long-term implications of life-saving but potentially non-restorative interventions are unknown early after injury. The neurological event usually occurs out of the blue and more than 80% of brain-injured patients are functionally normal before the event [6]. Resuscitation and other life-supporting measures are therefore initiated in order to maximise the chances of a good outcome whilst allowing time for a clearer assessment of prognosis [7]. Although there comes a time in all areas of medicine when continued treatment cannot benefit a patient, defining this point of no return after brain injury is enormously difficult and often impossible. Prognostic factors such as age, presenting Glasgow Coma Scale score, pupil reactivity and neuroimaging findings are well established but do not reliably predict individual mortality or the quality of outcome in survivors [8]. For example, there are specific factors, such as haematoma volume greater than 60 ml, hydrocephalus and intraventricular haemorrhage, that are well-accepted predictors of poor outcome after intracerebral haemorrhage [9], but some patients with several of these features may survive with only moderate degrees of disability [2]. Prognostic models, most commonly designed for traumatic brain injury, combine different characteristics of an individual patient to predict outcome more reliably, but also have limitations [10]. Progressive improvement can occur in some beyond one year but the degree of ultimate recovery thereafter is variable and unpredictable [11]. Cases of recovery after several years [12] are most likely to represent the exception rather than the norm. Thus, whilst it is important to reject therapeutic nihilism early after acute brain injury, it is equally important to maintain realistic outcome expectations. What constitutes a poor outcome, or an outcome in which continued aggressive therapy is futile, is a matter of much debate. Decisions involving futility of care are complex and, to a certain degree, dependent on individual judgments and values [13]. Pragmatically, treatment may be considered to be futile if it cannot improve a patient’s condition, but this is a subjective definition that implies a judgement on quality of life [14]. Decisions to withdraw life-sustaining treatment are usually made on the basis of a determination of an unfavourable prognosis for a meaningful neurological recovery [15]. In addition to the severity of illness and likely outcome [6], societal, cultural and religious pressures influence decisions regarding treatment limitation, and there are substantial differences in attitudes between and even within countries [16]. Some believe in the sanctity of life at all costs whereas others prefer treatment to be withdrawn if an acceptable functional outcome cannot be guaranteed [8]. Life-sustaining therapy is withheld or withdrawn after acute brain injury for two reasons – because it is believed either to be highly unlikely that a patient will survive, or that survival is likely but will be associated with an unacceptable (to the patient) quality of life. Because survival with severe disability is perceived by many to be an outcome worse than death [17], a ‘window of opportunity’ has been described in which decisions to withdraw treatment must be made [18]. The concern is that if this is deferred or delayed, the patient may no longer be physiologically dependent on intensive care treatments and there is a real risk of survival but with very severe impairment. The contrary view argues that this gives the impression that a patient risks missing the only ‘opportunity to die’ and that a fear of survival with unacceptable levels of disability can drive premature decisions to withdraw life-sustaining treatment [19]. ‘Early’ withdrawal of support is likely to result in a more rapid death, a lower risk of survival with severe disability and a lower societal burden, although at the risk of greater uncertainty about prognosis and insufficient time for considered decisions by surrogates. Conversely, ‘late’ withdrawal is associated with less uncertainty about prognosis and more time for surrogates to come to terms with the prognosis, but at the risk of a prolonged dying process and a higher risk of survival with severe disability. The challenge is to achieve a balance that maximises the chances of a good outcome for the maximum number of patients whilst minimising the possibility of producing large numbers of very severely disabled survivors. In appropriate patients, withdrawal of treatment should be seen as a key component of care rather than as a failure. Approximately 60% of deaths after acute brain injury are related to withdrawal of life-sustaining treatment [20]. Withdrawal of care is the most common immediate cause of death on neurointensive care units, usually in patients with devastating neurological injuries [6, 21]. Despite the uncertainty of prognostication soon after brain injury, a high proportion of deaths are related to early withdrawal of life-support. In a Canadian study, about one half of deaths that occurred following the withdrawal of treatment in patients with severe head injury occurred within the first three days [22]. Because it is the withdrawal of life support, rather than the brain injury itself, that often precipitates a patient’s death, treatment withdrawal may lead to a self-fulfilling prophecy. Prognosis is rarely absolute but the consequences of prognostic assessments often are [23]. In the absence of precise estimates of outcome after acute brain injury, it is often clinicians’ impressions and practice style that guide decisions to recommend withdrawal of life-support. A belief that survival with severe disability and/or impairment of cognitive function is the likely outcome is one of the strongest factors driving a recommendation to limit or withdraw treatment [24]. However, clinicians tend to be overly pessimistic when assessing prognosis [25]. In a prospective study in a medical intensive care unit, daily judgements of futility of treatment interventions with respect to survival and quality of life were made by doctors and nurses in 521 patients [26]. Although there was agreement in daily judgement in most cases, disagreement on at least one judgement was identified in 21% of patients overall and in 63% of dying patients. Disagreements were most common in the most severely ill patients and concerned perceptions of quality of life rather than survival. Nurses provided more pessimistic judgements and considered withdrawal of treatment more often than doctors, but the judgements of neither doctors nor nurses correctly predicted quality of life when compared with surviving patients’ own assessments of their status, with both falsely pessimistic and falsely optimistic judgements made by clinical staff. Similar pessimism is present during assessment of prognosis after brain injury; patients who are predicted to die using established prognostic factors can survive to hospital discharge and some may gain functional independence [2]. Multiple factors, including the specialty background of the treating physicians and the location of care, affect predictions of outcome and therefore the level of care provided [27]. Those who work regularly with critically ill brain-injured patients develop a deeper understanding of the factors that influence recovery, including the effects of brain plasticity and neurological rehabilitation, and apply more robust assessments of outcome and more realistic time-frames for recovery [28]. Limitation of care, whether it consists of withdrawal of life support [2], do-not-resuscitate orders [29] or both [30], is a major determinant of death after brain injury and a more powerful predictor of intensive care unit mortality than illness severity and systemic organ system dysfunction and failure [31]. Studies evaluating treatment and prognostic factors rarely take this into account and indicators that are used to identify poor prognosis are often determined by studies in which the presence of these same factors led to withdrawal of treatment rather than directly to death. This potentially biases predictive models and leads to self-fulfilling prophecies of poor outcome [2]. A further problem is that studies often use mortality as the primary outcome measure [2, 29, 30] whereas the most important issue for patients and families is whether survival is compatible with an acceptable quality of life [32]. Patients with severe brain injury are unable to make their own decisions, so when determining treatment options on their behalf the clinical team, in association with surrogate decision makers, must try and take into account what a patient’s preferences for continued treatment, including life-support, would be. Clinicians may consciously or unconsciously influence surrogate decision makers by imparting views that are primarily aligned with their personal beliefs and values, and confident prediction of poor neurological outcome, particular by doctors, may be subliminally, although unintentionally, coercive. It is also important to recognise that what an individual clinician considers a ‘good outcome’ might be highly challenging to patients and care-givers and, conversely, some patients might find acceptable an outcome deemed to be ‘poor’ by the treating clinical team. Judgements based on poorly informed or unfounded assumptions about a patient’s values and preferences must be avoided, and personal views about a patient’s likely quality of life should not impact on treatment decisions [7]. Many surrogates cite the patient’s known wishes as a ‘very important’ factor in deciding to withdraw support, and fewer than 2% base their decisions exclusively on prognostic information provided to them by doctors [15]. Poor communication with healthcare teams, including mixed messages and inconsistent estimates of prognosis, are often highlighted as sources of frustration by surrogate decision makers [33], and many believe that initial discussions about treatment withdrawal are premature [28]. Thus, if surrogates are ambivalent or insecure about a decision to withdraw life-support, it is best to continue treatment whilst these doubts are overcome. The option of organ donation should be an automatic consideration following withdrawal of treatment after acute brain injury [34]. It is crucially important that decisions to withdraw treatment are completely separated from consideration of organ donation but, once a consensus on treatment withdrawal has been reached, the local specialist nurse for organ donation should confirm the patient’s known wishes by review of the UK Organ Donor Register and, in association with local clinical staff, seek the views of surrogate decision makers [34]. Whilst there remains some resistance to donation after cardiac death from in some quarters, it is paternalistic to engage the family in decision-making on futility of treatment, and to withdraw treatment to accommodate the views and preferences of the patient, but not to offer the option of organ donation [34]. This approach is supported by the increasing emphasis placed on patient autonomy and endorsed in the Human Tissue Act. In most cases, death occurring beyond one hour after withdrawal of life-support leads to donor ineligibility because of potential organ compromise. Predicting the time to death after withdrawal of treatment is notoriously difficult but specific neurological and respiratory variables have been associated with earlier death after withdrawal of life-sustaining measures after brain injury, and the subset of patients with these features may be particularly suitable as potential donors after cardiac death [35]. Consideration of organ donation following a decision to withdraw life-sustaining treatment ensures that the wishes of all dying patients to donate their organs after death are identified, respected and honoured. There is considerable variability in current practice for withholding and withdrawing life-sustaining treatment after acute brain injury, reflecting different interpretations of outcome assessments as well as national, unit and personal biases [22]. High-quality research is needed to enable treatment-limiting decisions to be made within clinically relevant timescales and guided by improved prediction of outcomes that are meaningful to patients. Additional information from functional neuroimaging and other investigations might aid prognostic assessments in the future [36]. Decisions to withdraw treatment after acute brain injury are crucially important to the individual but also have wider societal implications, not least because the long-term financial and other resource burdens of caring for patients with severe disability are substantial. In a cash-limited healthcare system there is inevitable tension between the needs of the individual and those of the wider community [37], but it is for society, not doctors, to determine the personal, ethical and financial burden it is willing to place on individuals and on the community as a whole. MS is part funded by the Department of Health’s National Institute for Health Research Centres funding scheme via the University College London Hospitals/University College London Biomedical Research Centre. He was a member of the UK Organ Donation Taskforce.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.